Loom gives coding agents a shared index of your git repositories, served over MCP, so an
agent asks a structure question and gets the lines that answer it instead of reading whole
files. Needs a repository added and LOOM_CODE_INGEST_ENABLED=true. How the index is
built and kept fresh is in the code index overview;
what it gives in each language is in code index by language.
- Definitions, callers and impact.
code.symbol,code.callersandcode.impactlocate code with file and line references.code.index_statusreports each tier's indexed commit. Code operations - Search by meaning and by exact string.
code.searchfinds where a mechanism lives;code.grepfinds an exact message or identifier. - History and pull requests. Merged-PR history with review comments, a pre-PR
duplicate check and bisect candidates for a regression. These need PR history turned on
(
LOOM_CODE_PR_HISTORY_ENABLED); review comments and open PRs also need a GitHub token. History, PRs and releases says what each needs. - Lessons on the code.
code.symbolreturns the memories linked to a definition, andcode.search,code.callers,code.callees,code.impact,code.outlineand a task'scode.context_packattach the lessons for the files they return, each with whether its code moved, changed or is gone. Memory linked to code - Freshness you can check. Each tier can trail the branch. Inspect
as_ofwhen returned, or usecode.index_statusfor the relevant tier's watermark. Missing freshness metadata is unknown. Freshness - Plugins and MCP clients. The Claude Code plugin connects on the
codingtoolset and the Codex plugin oncore,coding,tracker,tasks; both list the code tools. Cursor, GitHub Copilot, Gemini CLI, Cline, Kilo Code, OpenCode and other MCP clients list them with?toolset=codingon their URL (loom installadds it for Cursor;--toolset codingadds it for the others it writes). Connect a coding agent · MCP clients
On the ClickHouse repository, Claude Code was asked to end each answer with up to three
files. With Loom it named the right file for 96.4% of 36 conceptual questions (10 paired
runs), against 91.4% without Loom, and the runs with Loom cost 21% less (the agent's model
bill; Loom's own model calls excluded).
On 200 SWE-bench Verified issues, code.search ranked a changed file in its first 10
results for 82% (file localization, not issue resolution), against 57% for BM25 over every
indexed file and 79% for BM25 over .py files only; the .py-only gap is within noise.
With and without Loom ·
SWE-bench Verified file localization
Start with a repository you trust. On the released stack, follow the operator-configured index steps; the dashboard's add-repository flow needs a sandbox that stack does not enable. Connect Claude Code or Codex with its plugin and start a session inside that repository's checkout (repository matching).
- Ask the agent to call
loom_opwithop: "code.index_status"andarguments: {"namespace": "<indexed namespace>", "brief": true}. Recordgit_head, the fast tier's indexed commit; an empty value means that tier has no watermark. - Pick a function whose definition you know. Ask the agent to call
loom_code_symbolin that namespace with the function's name andinclude_body: true. Check the returnedpath,start_line,end_line,bodyandtier. - Continue this check only for one unambiguous definition with
tier: "treesitter"and a complete, non-elided body. Compare its path, span and body with the source at the recordedgit_headbefore relying on it. Use a local read if the watermark or body is missing, the definition differs, or another tier is returned. Compiler tiers advance independently; their watermarks need a separate revision check (freshness). - Start a fresh session in the same checkout and repeat both calls. The index supplies the definition again; the agent need not load the previous chat or the whole file.
This checks one fast-tier definition. Uncommitted edits still need local reads, and a successful lookup is not a measurement of task accuracy or cost savings. See code operations for search, callers and impact.
Use the same indexed repository and choose a function with one direct caller you can identify in source. Its language needs call edges (language support). Replace the placeholders below with your organization, indexed namespace and qualified function name.
-
Call
loom_opwithop: "code.index_status"andarguments: {"org": "<org>", "namespace": "<indexed namespace>", "brief": false}. Recordgit_headand the returnedcompiler_tierscommits. Fast and compiler tiers can describe different revisions. -
Ask the agent to call
loom_opwith this request:{ "op": "code.callers", "arguments": { "org": "<org>", "namespace": "<indexed namespace>", "name": "<qualified function name>", "depth": 1, "max_nodes": 20, "include_ambiguous": false } }Check a returned
edgesrow'sfrom_symbol,path,line,edge_kind,tierandconfidence. Verify that the source at that tier's recorded commit contains the direct call; a reference to the name alone does not establish a call. Ifambiguousis true, choose the intended qualified name fromcandidatesfirst. -
Call
loom_opagain withop: "code.impact"andarguments: {"org": "<org>", "namespace": "<indexed namespace>", "name": "<qualified function name>", "depth": 1, "per_hop_limit": 20}. Look for the checked caller inimpacted_symbols:qualified_name,path,hop: 1andtier. This request resolves the supplied name;fileswould seed every function in those files.impacted_testslists mapped tests; the op does not execute them. -
Read
epistemic,causes,counts_floorandtruncatedon both answers, and repeat the status request. Compare returned locations with source at the relevant tier's commit. Missing fields or watermarks, changed watermarks, a missing expected caller or a source mismatch leave this check unverified (freshness).
This checks one caller's membership in the indexed graph. Counts remain floors,
including when epistemic is exact: runtime dispatch can be absent. Empty edges or
impact rows do not prove that a function has no callers or that a change is safe.
Adding repositories turns the index on for a repository; the code index overview explains the tiers.
Some ops on this page are marked experimental on the status page.