OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Developer Tools & SDKs

Vera

Vera is a Rust command-line tool for searching local code repositories. It combines BM25 lexical retrieval, vector similarity and optional cross-encoder reranking, parses code with tree-sitter, and returns ranked code chunks with file paths, line ranges and symbol metadata. Its documented capabilities include structural queries, reference lookup, project overviews, incremental indexing and an optional MCP server.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Vera — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Vera when you know what code does but not what it is called, or when an agent needs ranked, path-and-line-referenced code context. Use exact-string search such as ripgrep when you already know the literal text to find.

Who it is for

  • Developers searching repositories by behavior when they do not know the exact identifier.
  • AI-agent workflows that need focused, structured code context.
  • Teams seeking local code retrieval that can be exposed through an MCP server.

Documented uses

  • Find code related to a behavior or concept using both keyword and vector search.
  • Locate symbol-bounded code chunks and their file and line positions.
  • Run structural searches and look up references using the project index.
  • Refresh an existing index as project files change.
Before you choose

Requirements & limits

The project documents CLI installation through Bun, npm/npx or uvx. Local CPU is the default backend; API mode uses configured model endpoints and is described as paid usage. The repository is MIT-licensed.

  • Local CPU inference works offline after the initial model-asset download.
  • API mode sends code chunk text and queries to configured model endpoints.
  • Jina ONNX backends support optional local GPU inference.
  • The per-project index is stored in .vera/ and can be large.
  • Vera’s documented search coverage is more than 60 programming languages.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Vera, index a project and search locally

Set up the default local CPU backend, index the current project and retrieve relevant code with a search query.

Before you start

  • A supported environment for the Vera CLI.
  • Network access during the initial model-asset download.

Steps

  1. Install the CLI with a documented option such as bunx @vera-ai/cli install, the command below or uvx vera-ai install.

    Documented command
    npx -y @vera-ai/cli install

  2. Run vera setup --potion-code --index . in the project to download local model assets and index it.

  3. Add .vera/ to the project’s .gitignore.

  4. Run a query such as vera search "authentication logic".

Expected outcome

Vera returns ranked code chunks with file paths, line ranges and symbol metadata.

Guide limitations

  • The index can be large.
  • Initial model downloads require network access; local indexing and search work offline after download.
  • This guide covers the documented local CPU setup, not API-mode configuration.
Documentation 1 ↗
Source-based practical guide

Connect an MCP client to Vera

Configure a compatible MCP client to start Vera’s MCP server.

Before you start

  • Vera installed and available as the vera command.
  • An MCP client that supports command-based server configuration.

Steps

  1. Add a server entry with command vera and argument mcp to the client’s MCP configuration.

  2. For Claude Code, use the documented command claude mcp add vera -- vera mcp.

Expected outcome

The client can connect to Vera’s MCP server and use its documented tools, including `search_code`, `get_stats`, `get_overview`, `regex_search`, `structural_search`, `find_references` and `explain_path`.

Guide limitations

  • Client-specific setup details are maintained in Vera’s MCP integration documentation.
  • The repository lists tools and configuration examples; client compatibility depends on the MCP client.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
VeraTools/vera · Official homepage title · retained in its source language
How we review sourcesSuggest a correction