Updated the configuration templates to include the builtin preprocessors.

docs/source/distributions/remote_hosted_distro/nvidia.md

@@ -9,6 +9,7 @@ The `llamastack/distribution-nvidia` distribution consists of the following prov
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::nvidia` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/bedrock.md

@@ -16,6 +16,7 @@ The `llamastack/distribution-bedrock` distribution consists of the following pro
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::bedrock` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `remote::bedrock` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/cerebras.md

@@ -9,6 +9,7 @@ The `llamastack/distribution-cerebras` distribution consists of the following pr
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::cerebras`, `inline::sentence-transformers` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/fireworks.md

@@ -19,6 +19,7 @@ The `llamastack/distribution-fireworks` distribution consists of the following p
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::fireworks`, `inline::sentence-transformers` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/groq.md

@@ -19,6 +19,7 @@ The `llamastack/distribution-groq` distribution consists of the following provid
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::groq` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/meta-reference-gpu.md

@@ -19,6 +19,7 @@ The `llamastack/distribution-meta-reference-gpu` distribution consists of the fo
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `inline::meta-reference` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/meta-reference-quantized-gpu.md

@@ -19,6 +19,7 @@ The `llamastack/distribution-meta-reference-quantized-gpu` distribution consists
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `inline::meta-reference-quantized` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/remote-vllm.md

@@ -18,6 +18,7 @@ The `llamastack/distribution-remote-vllm` distribution consists of the following
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::vllm`, `inline::sentence-transformers` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/sambanova.md

@@ -17,6 +17,7 @@ The `llamastack/distribution-sambanova` distribution consists of the following p
 |-----|-------------|
 | agents | `inline::meta-reference` |
 | inference | `remote::sambanova` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | telemetry | `inline::meta-reference` |
 | tool_runtime | `remote::brave-search`, `remote::tavily-search`, `inline::code-interpreter`, `inline::rag-runtime` |

docs/source/distributions/self_hosted_distro/tgi.md

@@ -20,6 +20,7 @@ The `llamastack/distribution-tgi` distribution consists of the following provide
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::tgi`, `inline::sentence-transformers` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

docs/source/distributions/self_hosted_distro/together.md

@@ -19,6 +19,7 @@ The `llamastack/distribution-together` distribution consists of the following pr
 | datasetio | `remote::huggingface`, `inline::localfs` |
 | eval | `inline::meta-reference` |
 | inference | `remote::together`, `inline::sentence-transformers` |
+| preprocessing | `inline::basic`, `inline::simple_chunking` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |

llama_stack/templates/bedrock/bedrock.py

@@ -6,6 +6,7 @@
 
 from pathlib import Path
 
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import Provider, ToolGroupInput
 from llama_stack.providers.inline.vector_io.faiss.config import FaissVectorIOConfig
 from llama_stack.providers.remote.inference.bedrock.models import MODEL_ENTRIES
@@ -29,6 +30,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "bedrock"
     vector_io_provider = Provider(
@@ -57,6 +59,17 @@ def get_distribution_template() -> DistributionTemplate:
         ),
     ]
 
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
+
     return DistributionTemplate(
         name=name,
         distro_type="self_hosted",
@@ -72,6 +85,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=default_models,
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/bedrock/build.yaml

@@ -29,4 +29,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/bedrock/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -84,6 +85,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/bedrock}/registry.db
@@ -130,6 +138,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/cerebras/build.yaml

@@ -29,4 +29,7 @@ distribution_spec:
     - remote::tavily-search
     - inline::code-interpreter
     - inline::rag-runtime
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/cerebras/cerebras.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import ModelInput, Provider, ToolGroupInput
 from llama_stack.providers.inline.inference.sentence_transformers import (
     SentenceTransformersInferenceConfig,
@@ -33,6 +34,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::code-interpreter",
             "inline::rag-runtime",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
 
     name = "cerebras"
@@ -78,6 +80,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name="cerebras",
@@ -96,6 +108,7 @@ def get_distribution_template() -> DistributionTemplate:
                 default_models=default_models + [embedding_model],
                 default_shields=[],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/cerebras/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -86,6 +87,13 @@ providers:
   - provider_id: rag-runtime
     provider_type: inline::rag-runtime
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/cerebras}/registry.db
@@ -127,6 +135,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/ci-tests/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/ci-tests/ci_tests.py

@@ -6,6 +6,7 @@
 
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -38,6 +39,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "ci-tests"
     inference_provider = Provider(
@@ -70,6 +72,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
     available_models = {
         "fireworks": MODEL_ENTRIES,
     }
@@ -99,6 +111,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=default_models + [embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
                 default_shields=[ShieldInput(shield_id="meta-llama/Llama-Guard-3-8B")],
             ),
         },

llama_stack/templates/ci-tests/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -86,6 +87,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/ci-tests}/registry.db
@@ -215,6 +223,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/dell/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - remote::tavily-search
     - inline::code-interpreter
     - inline::rag-runtime
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/dell/dell.py

@@ -5,6 +5,7 @@
 # the root directory of this source tree.
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -33,6 +34,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::code-interpreter",
             "inline::rag-runtime",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "dell"
     inference_provider = Provider(
@@ -92,6 +94,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -107,6 +119,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
             "run-with-safety.yaml": RunConfigSettings(
                 provider_overrides={
@@ -120,6 +133,7 @@ def get_distribution_template() -> DistributionTemplate:
                 default_models=[inference_model, safety_model, embedding_model],
                 default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/dell/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -86,6 +87,13 @@ providers:
   - provider_id: rag-runtime
     provider_type: inline::rag-runtime
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/dell}/registry.db
@@ -116,6 +124,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/dell/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -82,6 +83,13 @@ providers:
   - provider_id: rag-runtime
     provider_type: inline::rag-runtime
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/dell}/registry.db
@@ -107,6 +115,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/dev/build.yaml

@@ -34,4 +34,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/dev/dev.py

@@ -7,6 +7,7 @@
 from typing import List, Tuple
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -91,6 +92,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "dev"
 
@@ -119,6 +121,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
     embedding_model = ModelInput(
         model_id="all-MiniLM-L6-v2",
         provider_id=embedding_provider.provider_id,
@@ -145,6 +157,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=default_models + [embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
                 default_shields=[ShieldInput(shield_id="meta-llama/Llama-Guard-3-8B")],
             ),
         },

llama_stack/templates/dev/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -103,6 +104,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/dev}/registry.db
@@ -359,6 +367,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/fireworks/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/fireworks/fireworks.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -39,6 +40,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
 
     name = "fireworks"
@@ -86,6 +88,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -104,6 +116,7 @@ def get_distribution_template() -> DistributionTemplate:
                 default_models=default_models + [embedding_model],
                 default_shields=[ShieldInput(shield_id="meta-llama/Llama-Guard-3-8B")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
             "run-with-safety.yaml": RunConfigSettings(
                 provider_overrides={
@@ -149,6 +162,7 @@ def get_distribution_template() -> DistributionTemplate:
                     ),
                 ],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/fireworks/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -95,6 +96,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/fireworks}/registry.db
@@ -229,6 +237,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/fireworks/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -89,6 +90,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/fireworks}/registry.db
@@ -218,6 +226,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/groq/build.yaml

@@ -26,4 +26,7 @@ distribution_spec:
     - remote::tavily-search
     - inline::code-interpreter
     - inline::rag-runtime
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/groq/groq.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -36,6 +37,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::code-interpreter",
             "inline::rag-runtime",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "groq"
 
@@ -77,6 +79,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -93,6 +105,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=default_models + [embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/groq/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -86,6 +87,13 @@ providers:
   - provider_id: rag-runtime
     provider_type: inline::rag-runtime
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/groq}/registry.db
@@ -152,6 +160,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/hf-endpoint/build.yaml

@@ -29,4 +29,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/hf-endpoint/hf_endpoint.py

@@ -5,6 +5,7 @@
 # the root directory of this source tree.
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -36,6 +37,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "hf-endpoint"
     inference_provider = Provider(
@@ -84,6 +86,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -100,6 +112,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
             "run-with-safety.yaml": RunConfigSettings(
                 provider_overrides={
@@ -123,6 +136,7 @@ def get_distribution_template() -> DistributionTemplate:
                 ],
                 default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/hf-endpoint/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -94,6 +95,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-endpoint}/registry.db
@@ -124,6 +132,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/hf-endpoint/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -89,6 +90,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-endpoint}/registry.db
@@ -114,6 +122,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/hf-serverless/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/hf-serverless/hf_serverless.py

@@ -5,6 +5,7 @@
 # the root directory of this source tree.
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -36,6 +37,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
 
     name = "hf-serverless"
@@ -85,6 +87,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -101,6 +113,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
             "run-with-safety.yaml": RunConfigSettings(
                 provider_overrides={
@@ -124,6 +137,7 @@ def get_distribution_template() -> DistributionTemplate:
                 ],
                 default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/hf-serverless/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -94,6 +95,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-serverless}/registry.db
@@ -124,6 +132,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/hf-serverless/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -89,6 +90,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-serverless}/registry.db
@@ -114,6 +122,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/meta-reference-gpu/build.yaml

@@ -29,4 +29,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/meta-reference-gpu/meta_reference.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -40,6 +41,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "meta-reference-gpu"
     inference_provider = Provider(
@@ -91,6 +93,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -106,6 +118,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
             "run-with-safety.yaml": RunConfigSettings(
                 provider_overrides={
@@ -130,6 +143,7 @@ def get_distribution_template() -> DistributionTemplate:
                 ],
                 default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/meta-reference-gpu/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -96,6 +97,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-reference-gpu}/registry.db
@@ -126,6 +134,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/meta-reference-gpu/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -90,6 +91,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-reference-gpu}/registry.db
@@ -115,6 +123,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/meta-reference-quantized-gpu/build.yaml

@@ -29,4 +29,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/meta-reference-quantized-gpu/meta_reference.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import ModelInput, Provider, ToolGroupInput
 from llama_stack.providers.inline.inference.meta_reference import (
     MetaReferenceQuantizedInferenceConfig,
@@ -35,6 +36,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     default_tool_groups = [
         ToolGroupInput(
@@ -50,6 +52,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
     name = "meta-reference-quantized-gpu"
     inference_provider = Provider(
         provider_id="meta-reference-inference",
@@ -96,6 +108,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/meta-reference-quantized-gpu/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -92,6 +93,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-reference-quantized-gpu}/registry.db
@@ -117,6 +125,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/nvidia/build.yaml

@@ -27,4 +27,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/nvidia/nvidia.py

@@ -6,6 +6,7 @@
 
 from pathlib import Path
 
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import Provider, ToolGroupInput
 from llama_stack.providers.remote.inference.nvidia import NVIDIAConfig
 from llama_stack.providers.remote.inference.nvidia.models import MODEL_ENTRIES
@@ -29,6 +30,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
 
     inference_provider = Provider(
@@ -54,6 +56,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     default_models = get_model_registry(available_models)
     return DistributionTemplate(
@@ -71,6 +83,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=default_models,
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/nvidia/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -86,6 +87,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/nvidia}/registry.db
@@ -220,6 +228,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/passthrough/build.yaml

@@ -29,4 +29,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/passthrough/run.yaml

@@ -89,6 +89,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-llama}/registry.db
@@ -116,5 +123,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/remote-vllm/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/remote-vllm/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -96,6 +97,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/registry.db
@@ -126,6 +134,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/remote-vllm/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -90,6 +91,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/registry.db
@@ -115,6 +123,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/remote-vllm/vllm.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -38,6 +39,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "remote-vllm"
     inference_provider = Provider(
@@ -88,6 +90,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -103,6 +115,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
             "run-with-safety.yaml": RunConfigSettings(
                 provider_overrides={
@@ -126,6 +139,7 @@ def get_distribution_template() -> DistributionTemplate:
                 ],
                 default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/sambanova/build.yaml

@@ -19,4 +19,7 @@ distribution_spec:
     - remote::tavily-search
     - inline::code-interpreter
     - inline::rag-runtime
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/sambanova/run.yaml

@@ -3,6 +3,7 @@ image_name: sambanova
 apis:
 - agents
 - inference
+- preprocessing
 - safety
 - telemetry
 - tool_runtime
@@ -64,6 +65,13 @@ providers:
   - provider_id: rag-runtime
     provider_type: inline::rag-runtime
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/sambanova}/registry.db
@@ -171,6 +179,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/sambanova/sambanova.py

@@ -6,6 +6,7 @@
 
 from pathlib import Path
 
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     Provider,
     ShieldInput,
@@ -29,6 +30,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::code-interpreter",
             "inline::rag-runtime",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "sambanova"
 
@@ -56,6 +58,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -73,6 +85,7 @@ def get_distribution_template() -> DistributionTemplate:
                 default_models=default_models,
                 default_shields=[ShieldInput(shield_id="meta-llama/Llama-Guard-3-8B")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/tgi/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/tgi/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -89,6 +90,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/tgi}/registry.db
@@ -114,6 +122,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/tgi/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -88,6 +89,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/tgi}/registry.db
@@ -113,6 +121,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/tgi/tgi.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -38,6 +39,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "tgi"
     inference_provider = Provider(
@@ -88,6 +90,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -104,6 +116,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
             "run-with-safety.yaml": RunConfigSettings(
                 provider_overrides={
@@ -125,6 +138,7 @@ def get_distribution_template() -> DistributionTemplate:
                 ],
                 default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}")],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/together/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/together/run-with-safety.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -95,6 +96,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/together}/registry.db
@@ -226,6 +234,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/together/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -89,6 +90,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/together}/registry.db
@@ -215,6 +223,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/together/together.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import (
     ModelInput,
     Provider,
@@ -39,6 +40,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
     name = "together"
     inference_provider = Provider(
@@ -74,6 +76,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
     embedding_model = ModelInput(
         model_id="all-MiniLM-L6-v2",
         provider_id="sentence-transformers",
@@ -99,6 +111,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=default_models + [embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
                 default_shields=[ShieldInput(shield_id="meta-llama/Llama-Guard-3-8B")],
             ),
             "run-with-safety.yaml": RunConfigSettings(
@@ -145,6 +158,7 @@ def get_distribution_template() -> DistributionTemplate:
                     ),
                 ],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={

llama_stack/templates/vllm-gpu/build.yaml

@@ -30,4 +30,7 @@ distribution_spec:
     - inline::code-interpreter
     - inline::rag-runtime
     - remote::model-context-protocol
+    preprocessing:
+    - inline::basic
+    - inline::simple_chunking
 image_type: conda

llama_stack/templates/vllm-gpu/run.yaml

@@ -5,6 +5,7 @@ apis:
 - datasetio
 - eval
 - inference
+- preprocessing
 - safety
 - scoring
 - telemetry
@@ -92,6 +93,13 @@ providers:
   - provider_id: model-context-protocol
     provider_type: remote::model-context-protocol
     config: {}
+  preprocessing:
+  - provider_id: basic
+    provider_type: inline::basic
+    config: {}
+  - provider_id: simple_chunking
+    provider_type: inline::simple_chunking
+    config: {}
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/vllm-gpu}/registry.db
@@ -117,6 +125,10 @@ tool_groups:
   provider_id: rag-runtime
 - toolgroup_id: builtin::code_interpreter
   provider_id: code-interpreter
-preprocessors: []
+preprocessors:
+- preprocessor_id: builtin::basic
+  provider_id: basic
+- preprocessor_id: builtin::chunking
+  provider_id: simple_chunking
 server:
   port: 8321

llama_stack/templates/vllm-gpu/vllm.py

@@ -5,6 +5,7 @@
 # the root directory of this source tree.
 
 from llama_stack.apis.models.models import ModelType
+from llama_stack.apis.preprocessing.preprocessors import PreprocessorInput
 from llama_stack.distribution.datatypes import ModelInput, Provider
 from llama_stack.providers.inline.inference.sentence_transformers import (
     SentenceTransformersInferenceConfig,
@@ -35,6 +36,7 @@ def get_distribution_template() -> DistributionTemplate:
             "inline::rag-runtime",
             "remote::model-context-protocol",
         ],
+        "preprocessing": ["inline::basic", "inline::simple_chunking"],
     }
 
     name = "vllm-gpu"
@@ -80,6 +82,16 @@ def get_distribution_template() -> DistributionTemplate:
             provider_id="code-interpreter",
         ),
     ]
+    default_preprocessors = [
+        PreprocessorInput(
+            preprocessor_id="builtin::basic",
+            provider_id="basic",
+        ),
+        PreprocessorInput(
+            preprocessor_id="builtin::chunking",
+            provider_id="simple_chunking",
+        ),
+    ]
 
     return DistributionTemplate(
         name=name,
@@ -96,6 +108,7 @@ def get_distribution_template() -> DistributionTemplate:
                 },
                 default_models=[inference_model, embedding_model],
                 default_tool_groups=default_tool_groups,
+                default_preprocessors=default_preprocessors,
             ),
         },
         run_config_env_vars={