ollama_corpus_index

Index corpus. Build a persistent named corpus. Pass name + paths: string[]; the server chunks, embeds, stores the corpus at ~/.ollama-intern/corpora/.json AND writes a manifest at .manifest.json that captures the declared paths + chunk params + embed model. Idempotent — unchanged files are reused by sha256; changed files are re-embedded; paths not in the input are dropped. For day-to-day upkeep once a manifest exists, prefer ollama_corpus_refresh — it reconciles corpus vs manifest and reports drift.

Parameters

Parameter	Type	Required	Default	Description
name	string	yes	—	Corpus name (e.g. ‘memory’, ‘canon’, ‘handbook’). Maps to a file under ~/.ollama-intern/corpora/.
paths	string[]	yes	—	Absolute file paths to include in the corpus. Pass an explicit list of files — corpus_index does NOT walk directories; you enumerate. Unchanged files are reused from the existing corpus by sha256.
chunk_chars	integer	no	—	Chars per chunk (default 800).
chunk_overlap	integer	no	—	Overlap between adjacent chunks (default 100).

Full input schema

The JSON Schema below is generated from the same zod schema the server validates against at the wire — it cannot drift from runtime behavior.

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "type": "object",
  "properties": {
    "name": {
      "type": "string",
      "minLength": 1,
      "pattern": "^[a-zA-Z0-9_-]+$",
      "description": "Corpus name (e.g. 'memory', 'canon', 'handbook'). Maps to a file under ~/.ollama-intern/corpora/."
    },
    "paths": {
      "minItems": 1,
      "type": "array",
      "items": {
        "type": "string",
        "minLength": 1
      },
      "description": "Absolute file paths to include in the corpus. Pass an explicit list of files — corpus_index does NOT walk directories; you enumerate. Unchanged files are reused from the existing corpus by sha256."
    },
    "chunk_chars": {
      "description": "Chars per chunk (default 800).",
      "type": "integer",
      "minimum": 100,
      "maximum": 8000
    },
    "chunk_overlap": {
      "description": "Overlap between adjacent chunks (default 100).",
      "type": "integer",
      "minimum": 0,
      "maximum": 4000
    }
  },
  "required": [
    "name",
    "paths"
  ]
}

Envelope

Every tool returns the standard envelope — tier_used, model, tokens_in/tokens_out, elapsed_ms, backend provenance, warnings. See Envelope & tiers.

Source

Schema + handler: src/tools/corpusIndex.ts · registration: src/index.ts

---

Back to the Tool Reference.