sqry

Overview

sqry is a locally run code-analysis tool for developers who need to understand how symbols relate across a codebase. It is best suited to maintainers who want structural queries in an editor or AI-assisted workflow; the need to index a project first makes it less compelling for a quick, one-off lookup.

Instead of relying solely on text matches, sqry parses code into an abstract syntax tree and builds a graph of symbols and relationships. Its index persists, so later queries can use the cached graph snapshot. Regular-expression search and code navigation complement the structural analysis.

Key features

Queries and language support

Documented queries and graph commands cover callers, callees, implementations, dependency cycles, and unused symbols. That makes sqry useful for tracing behavior and examining code structure, while the language support has an important distinction: of 37 plugins, 28 provide full relation tracking and nine extract symbols and imports. Users should confirm their language has the relationship support needed for their queries.

Editor and AI integrations

The LSP server works with VS Code, Neovim, Helix, and any LSP 3.17 client, putting navigation in established editor workflows. For AI-assisted work, the MCP server supports Claude, Codex, Gemini, Cursor, and Windsurf. Its standalone mode exposes 39 tools; daemon mode offers a curated subset of 16, which is a more constrained interface for integrations that do not need the full set.

Local operation and setup

sqry runs on the user's machine under the MIT license. Its maker says source code stays on the machine and the software has no telemetry, which suits users who want local processing. The trade-off is that users must build an index before querying, even though it persists for later use.

Installers are provided for macOS, Linux, and Windows, and cargo installation is available on all three. Windows users need the bundled MinGW runtime DLLs alongside the executables; the release ZIP includes them. Public releases include checksums and Sigstore DSSE attestation metadata, and installation scripts verify checksums by default.

CI workflows

The documentation describes running the install script and CLI commands in GitHub Actions, and the home page identifies a GitHub Action for checking dead code, cycles, and duplicates on pull requests. Teams should expect to wire up CI using those approaches: sqry does not currently publish a public composite action.

Pricing

sqry is free: its sqry plan costs nothing and includes the open-source MIT-licensed software running locally. There is no paid tier or trial to weigh against it. The free plan's central limitation is practical rather than a stated usage cap: users need to build an index, and the depth of relationship queries depends on the language plugin.

Platforms

sqry supports Linux, macOS, and Windows. The local deployment model and installers make it a fit for developers working on those desktop platforms; Windows setup requires keeping the included MinGW runtime DLLs with the executables.

Who it's for

sqry is a strong fit for developers and maintainers who need to follow calls, implementations, dependencies, or unused symbols, especially when they want those capabilities in an editor or through an MCP-connected assistant. It is also relevant to teams seeking local code processing or a CLI-based GitHub Actions workflow. It is a weaker fit for users who only need immediate text search, or whose language depends on relationship tracking when its plugin provides only symbol and import extraction.

Pros and cons

• Pros: Structural queries span callers, callees, implementations, cycles, and unused symbols, offering more than regular-expression matching alone.
• Pros: LSP and MCP integrations connect the tool to named editors and AI assistants, with a choice between 39 standalone tools and a 16-tool daemon subset.
• Pros: Local execution, MIT licensing, and the maker's no-telemetry and no-source-transfer claims suit privacy-conscious development workflows.
• Cons: Indexing is required before the first query, adding a setup step for one-off searches.
• Cons: Nine of 37 language plugins do not provide full relation tracking, which can rule out some structural queries for those languages.
• Cons: CI users must rely on the install script and CLI commands rather than a public composite action.

Alternatives

For a different code-search workflow, compare the Code Search Tools category. Choose ast-grep if you want a free tool available across API, extension, desktop, and web platforms. Code Search is a better fit for teams seeking unlimited users, full-text search, repository-wide find and replace, and CLI and MCP access on its free Community plan.

OpenGrok is another free, open-source option, under the CDDL license. Choose Shebe if local code search under the Apache-2.0 license is the priority. Chatterbox TTS is an alternative with free, MIT-licensed open-source models for research and development.

LXR Cross Referencer may suit users looking for free, self-installed, open-source software. Semantic Designs Source Code Search Engine offers a free evaluation version limited to 10,000 lines of Java, COBOL, C++, and pseudo-code. Sourcebot is a freemium option with a free Basic plan for code search, including regex and symbol search.

Verdict

Choose sqry if you need local, graph-based code queries and want to bring them into an editor or AI assistant without paying for an MIT-licensed tool. Its most important caveat is uneven relationship support across language plugins, so look elsewhere if your language needs a level of tracking its plugin does not provide or you want search without an indexing step.