Silica Core

Lightweight, local evidence retrieval tools for code and research

Silica locates the source, symbol, page, or passage you and your agents need. Maximum signal, minimum machinery.

Silica indexes the markdown, code, PDFs and office files under one root and serves them to Claude Code, Cursor, Hermes, Codex, OpenCode and every other popular harness, as MCP tools or as shell commands that print the same JSON. A hit is a path, a section, a line, a window of text and the numbers to judge it by. The harness owns the loop.

the quickstart recorded end to end: uv tool install, silica init reporting nine indexed documents, silica setup claude registering the MCP server, then Claude Code answering a question about the LSM compaction design space by calling silica-core and citing the PDF with its page and its line

Three commands, then a question asked the way you would ask any other. The harness calls silica_search, and the answer carries the file, the page and the line it came from.

Install

uv tool install 'silica-core[mcp]'         # BM25 over documents and their sections (faster, lighter)
uv tool install 'silica-core[mcp,dense]'   # in addition the dense leg: numpy, a static model (still fast, more precise)

The second line adds the dense leg: numpy and a static model2vec model, no torch and no GPU. It stays inert until the model is named and the sections are embedded — the last stanza of the Quickstart. Take it when the questions are paraphrases that share no words with the text; an exact term or an identifier is answered by the lexical leg either way, and only that leg reports terms_absent.

pipx works the same. The package is silica-core, the command is silica, the tools are silica_*.

Quickstart

In any folder of markdown, code, PDFs or office files:

silica init                               # adopt the folder: ignore file, vault.yaml, first index; the MCP server serves adopted folders only
silica search "leveled compaction" -k 5  # the best located passages
silica setup claude                      # register the MCP server, write the guidance block into ~/.claude/CLAUDE.md

# optional, with the [dense] extra: the dense leg, a static model, nothing leaves the machine
export SILICA_EMBEDDING_MODEL=model2vec/minishlab/potion-retrieval-32M
silica index --embed

Nine arXiv papers, indexed in 2.1 s. The hit names the file, the page and the passage; the page beside it is the check.

Tools

In a source tree every function, method, class and constant is its own unit: a hit's section is the symbol, span its lines, and silica_read(path, section=…) serves the body. For a symbol whose name is known, grep wins; for a question that names none, the search comes first: the tool description and the block silica setup claude writes say so. The contract, the reply shapes and the acceptance checks are in TOOLS.md. Nothing needs an API key or a network.

How search says no

A ranked list always has a top, even when the corpus does not answer. Three fields say how much the result is worth:

• coverage: the share of the query's idf mass the hit's matched terms carry. Near 1, every rare term matched; near 0, only common words did.
• terms_absent: query terms that occur nowhere in the corpus.
• matched_terms: the words this hit actually contains.

raft consensus log replication over the same nine papers: raft and consensus occur in none of them, coverage falls to 0.19, and the top hit is about data replication. On 254 papers the top hit of an answered question carries 0.69 to 1.00; a question the corpus does not cover, 0.44. Silica exposes the signals; the harness decides whether to stop, read or rephrase.

Benchmarks

nDCG@10 on BEIR SciFact · NFCorpus, the same documents and queries for every arm. BEIR's published BM25 baselines are 0.665 · 0.325. Silica's lexical index needs no model; the others serve lexical search from an index that also holds embeddings.

Code, on the twenty SWE-QA questions zvec-grep publishes for its own benchmark, same embedder, k = 10, scored on the files and symbols the reference answer rests on.

On BEIR the two hybrids tie at the 95% interval: the same vectors rank the same, with no daemon and no vector store. On code, Silica's hybrid file MRR is +0.135 over zvec-grep's hybrid (95% interval +0.01 to +0.27), paired per question. The fusion also gains +0.21 MRR and +0.30 symbol hit over Silica's lexical arm; no reranker or graph expansion is involved.

On this measured scope, Silica is a compact, local, SOTA-competitive retriever: it matches zvec-grep on BEIR and leads the paired SWE-QA code-localization replay with the same embedder.

Retrieval matters only if the agent does less work without losing the answer. These are separate experiments and are not pooled:

The repository result is one repetition: turns improve by 1.75 (95% interval 0.55 to 3.05 fewer), while Judge and cost remain inconclusive. The document rows belong to a 144-run study over twelve questions and a 5.5M-token corpus. They establish less work on that workload, not a universal agent claim.

Corpora, intervals, per-task exceptions and reproduction commands are in benchmarks.

Harnesses

silica setup <client> writes the registration into the client's own config and backs up what was there; for claude it also puts a guidance block, when to search before grep, into ~/.claude/CLAUDE.md. silica setup --list names the clients: claude, codex, cursor, windsurf, zed, cline, roo, continue, goose, opencode, openhands, gemini, dsh, hermes, openclaw, agent-zero, claude-desktop, lmstudio, anythingllm and librechat; shell, python and generic print recipes for anything else. The server serves the folder the client opens in; --vault DIR or SILICA_VAULT fixes the root.

Every written block, and the Claude Code plugin, run silica mcp --retrieval local-hybrid: potion-retrieval-32M in the server process, index and vectors built in the background at start, the search lexical and dense: warming until they land. Nothing leaves the machine; the one download is the model, once. npx skills add kiycoh/silica-core installs the skill that tells an agent when to reach for the tools, and nothing else. Shell recipes, Docker and the REPL are in public/harnesses.md.

Notes

• What it reads: markdown, .txt, .rst and PDFs with a text layer directly, one PDF page per section; DOCX, EPUB, FB2, RTF, XLS and ODF converted with no extra; scanned PDFs, images, PPTX and XLSX through silica import with MinerU; audio and video with ffmpeg plus SILICA_STT_BASE_URL; CSV readable by line, excluded from search. In a source tree the code lane adds source files and their json, yaml, toml, cfg and ini. silica doctor says which lanes this machine has.
• The dense leg: section embeddings that catch a paraphrase sharing no rare word with the answer. uv tool install stays lexical until silica index --embed, with the [dense] extra's local model or any OpenAI-compatible /v1/embeddings endpoint. Text leaves the machine only for a remote endpoint, and only after silica index --embed --allow-remote grants that host once. Variables and reply states in TOOLS.md, the ollama recipe in public/harnesses.md.
• More surfaces: silica mcp --extended adds the wikilink tools; silica connect (extra [connect]) hosts the bridge the Obsidian plugin dials into, so writes land through the vault API while the app is open; silica repl runs a small reference agent over the same tools, the one surface that needs a model (SILICA_MODEL).
• Not in the core: no memory lane, prompt injection, summaries or undo journal. Undo is git.

License

MIT. See LICENSE.