ollama_code_citation

Code citations. RESEARCH. Answer a code question with PER-CLAIM CITATIONS (file + line range). Distinct from ollama_research: research cites files, code_citation cites LINES. Pass question (10-1000 chars), source_paths[], optional per_file_max_chars (default 100_000). Server numbers each line 1-based in the prompt so the Deep tier can anchor claims exactly. Returns {answer, citations:[{claim_fragment, file, start_line, end_line, excerpt}], uncited_fragments[], weak}. Citations to files outside source_paths are stripped (same rule as ollama_research). Citations with line ranges outside the loaded file bounds are also stripped — each stripping reason lands in warnings[]. Empty answer or answer-without-citations flips weak=true.

Parameters

Parameter	Type	Required	Default	Description
question	string	yes	—	The code question to answer. 10-1000 chars — specific enough that citations can be anchored (‘where does X get validated’, not ‘tell me about this code’).
source_paths	any	yes	—	Files the answer must be grounded in. Citations to files outside this list are stripped.
per_file_max_chars	integer	no	—	Chars to read per file (default 100_000). Bigger than research because citations need the lines to exist.
model	string	no	—	Optional per-call model override. When provided, overrides the tool’s tier-resolved model for this call. The tier’s timeout (TIER_TIMEOUT_MS) still applies. On timeout, fallback uses the tier-resolved model, NOT the override. Use for receipt-backed orchestration that requires explicit model identity (e.g., research-os reviewer profiles).

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": {
    "question": {
      "type": "string",
      "minLength": 10,
      "maxLength": 1000,
      "description": "The code question to answer. 10-1000 chars — specific enough that citations can be anchored ('where does X get validated', not 'tell me about this code')."
    },
    "source_paths": {
      "description": "Files the answer must be grounded in. Citations to files outside this list are stripped."
    },
    "per_file_max_chars": {
      "description": "Chars to read per file (default 100_000). Bigger than research because citations need the lines to exist.",
      "type": "integer",
      "minimum": 1000,
      "maximum": 500000
    },
    "model": {
      "description": "Optional per-call model override. When provided, overrides the tool's tier-resolved model for this call. The tier's timeout (TIER_TIMEOUT_MS) still applies. On timeout, fallback uses the tier-resolved model, NOT the override. Use for receipt-backed orchestration that requires explicit model identity (e.g., research-os reviewer profiles).",
      "type": "string",
      "minLength": 1
    }
  },
  "required": [
    "question",
    "source_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/codeCitation.ts · registration: src/index.ts

---

Back to the Tool Reference.