Laya "System 1" Code Intelligence for AI Agents

Laya released SCS 0.2.0, a headless code-intelligence service that indexes source repositories and lets coding agents query them through a local MCP tool called query_code. SCS starts with an empty index, enrolls a repository only after an explicit CLI, MCP, or client request, and never modifies repository source; stable releases support Apple Silicon macOS and x86-64 Linux with CPython 3.14. Indexing requires an embedding provider, defaulting to the OpenAI embeddings API with text-embedding-3-large at 3072 dimensions, while Laya is an optional local Apple Silicon feature for selecting among seven query playbooks and SCS works without it on both supported platforms.

SCS is a headless code-intelligence service. It indexes source repositories and lets coding agents investigate them through a local MCP tool. An agent asks query code a question; SCS returns bounded evidence from its structural and semantic index. SCS starts with an empty index. It enrolls a repository only after an explicit CLI, MCP, or client request, and never changes repository source. - Stable releases support Apple Silicon macOS and x86-64 Linux with CPython 3.14. - Indexing needs an embedding provider. The default uses the OpenAI embeddings API and sends source-derived entity text to it. Local providers are available. - Laya is an optional, local Apple Silicon feature for choosing query playbooks. SCS works without Laya on both supported platforms. See Laya requirements optional-laya-routing for its measured memory use. Download the installer and checksum manifest from the same GitHub Release https://github.com/leonardoventurini/scs/releases , verify the installer, and install SCS: VERSION=0.2.0 curl -fsSLO "https://github.com/leonardoventurini/scs/releases/download/v${VERSION}/scs-installer-${VERSION}.sh" curl -fsSLO "https://github.com/leonardoventurini/scs/releases/download/v${VERSION}/SHA256SUMS" shasum -a 256 -c SHA256SUMS --ignore-missing sh "scs-installer-${VERSION}.sh" scs version On Linux, use sha256sum -c SHA256SUMS --ignore-missing . The installer verifies its wheel and constraints, installs without sudo , and uses a pinned, checksum-verified uv binary when necessary. Current macOS releases are not Apple-signed or notarized. See distribution and upgrade details https://github.com/leonardoventurini/scs/blob/main/docs/github-releases-distribution.md . Configure an embedding provider before indexing. For the default OpenAI provider, put this in ~/.scs/config.toml : embedding provider = "openai" embedding model = "text-embedding-3-large" embedding dimension = 3072 openai api key = "replace-with-your-key" Keep the file owner-readable only chmod 600 ~/.scs/config.toml . See embedding configuration https://github.com/leonardoventurini/scs/blob/main/docs/configuration.md for local provider and reranking options. Register the installed stdio bridge with Codex, then index the repository containing your current directory: codex mcp add scs -- "$HOME/.local/bin/scs" mcp codex mcp get scs scs index "$PWD" scs status If an existing scs MCP entry points elsewhere, remove it first with codex mcp remove scs . Restart open Codex clients after changing MCP configuration. Indexing runs as a durable background job; use scs status or get graph stats to check when it is ready. The path above assumes the installer's default ~/.local/bin location. agent goal + repository + optional anchors | v validate request and paths | v select one of seven playbooks <--- optional local Laya classifier | v bounded index search and graph reads | v evidence + routing + trace + completeness For example, an agent can ask: query code goal="Find tests affected by changes to the parser", repo path="/repo", file paths= "src/parser.py" , mode="balanced", The fast , balanced , and thorough modes set fixed time and evidence budgets. Results show which playbook ran and whether evidence was complete, truncated, or degraded. See the MCP tool reference https://github.com/leonardoventurini/scs/blob/main/docs/mcp-tools.md for anchors, tool contracts, and the query code migration guide https://github.com/leonardoventurini/scs/blob/main/docs/query-code-migration.md for retired tools. SCS indexes supported source files structurally. Other regular UTF-8 text files can be indexed at file level for lexical and semantic search. Git ignore rules and size limits apply. Once a repository is enrolled, SCS watches Git-visible changes and updates its index in the background. See indexing and project management https://github.com/leonardoventurini/scs/blob/main/docs/indexing.md for coverage, limits, reindexing, and deletion. The default OpenAI embedding provider sends source-derived entity text to the configured API. SCS does not send whole repository files to a summarization service. You can instead configure a local OpenAI-compatible server or an in-process MLX provider. Provider details and trust controls are in embedding configuration https://github.com/leonardoventurini/scs/blob/main/docs/configuration.md . On Apple Silicon, Laya can choose one of SCS's bounded query playbooks. SCS runs Laya in its own local MLX worker process; it does not call an external inference service. Laya receives the goal and explicit anchors, not repository source, embeddings, or retrieved evidence. SCS performs the search and graph reads. Without Laya, routing follows deterministic rules. Resource example: On a Mac Studio M3 Ultra, the pinned model bundle occupied about 807 MB on disk, and a warmed Laya worker measured about 5.2 GB of physical memory footprint on 2026-09-24. This is one observed measurement, not a fixed minimum; usage can vary by host and workload. Laya is disabled unless explicitly configured. To enable it from a source checkout on Apple Silicon: uv sync --all-groups --extra laya uv run --extra laya python scripts/install-laya.py Add decision model = "laya" to ~/.scs/config.toml , then run uv run --extra laya scs daemon restart from that checkout. The release installer installs the base tool without the optional Laya dependency. The installation script downloads and verifies a pinned model bundle; queries never download a model. A configured daemon reports ready only after its worker loads and warms. If inference fails during a query, SCS reports degradation and uses deterministic routing. scs list shows enrolled projects and their stable numeric IDs. scs reingest ID|PATH forces a full rebuild, and scs delete ID|PATH removes only SCS-owned derived state. scs doctor checks daemon health; scs metrics --days 7 --json reports aggregate operations without query text, source text, file paths, job payloads, or results. See indexing and project management https://github.com/leonardoventurini/scs/blob/main/docs/indexing.md for lifecycle details. Each MCP client runs a small stdio bridge. Bridges share one lazily started daemon, which shuts down after the last bridge disconnects. Persistent state lives under SCS HOME . See architecture https://github.com/leonardoventurini/scs/blob/main/docs/architecture.md for storage, runtime ownership, and legacy-index migration. For a source checkout: just setup just verify just eval-search just setup syncs dependencies, builds the private native extension, and installs the repository's pre-commit hook. just verify runs strict Basedpyright checks, Ruff, Python tests with branch coverage, and the Rust workspace tests. Search and query evaluation guidance lives in evals/README.md https://github.com/leonardoventurini/scs/blob/main/evals/README.md .