Introduction to Sourcegraph

Sourcegraph is a code intelligence platform used by engineering teams at companies ranging from a few hundred developers to more than 40,000. It indexes all of your repositories across every code host, giving developers and engineering leaders a complete, up-to-date understanding of their codebase.

​

Key features

​

Code Search

Sourcegraph code search is fast, works across all your repositories at any commit, and has minimal indexing delay. Search across every branch and every code host simultaneously using powerful query syntax. Key capabilities include:

• Regex and structural search — match patterns, code blocks, and nested expressions using Comby syntax
• Commit diff and message search — find changes to particular functions, classes, or areas of the codebase over time
• Multi-line search — search queries that cross multiple lines, not just line-oriented patterns
• Search contexts — scope searches to a curated set of repositories and revisions
• Symbol search — filter results to functions, variables, and other symbols across 75+ languages

​

Cody AI

Cody is Sourcegraph’s AI coding assistant. Unlike general-purpose AI tools, Cody has full context over your entire codebase—not just the files open in your editor—making its answers more accurate and grounded.

• Chat — ask questions about your codebase and get sourced, detailed answers
• Autocomplete — context-aware inline code completions as you type
• Commands — built-in actions for explaining code, generating tests, fixing bugs, and more
• Multiple clients — available in VS Code, JetBrains IDEs, and the Sourcegraph web UI

​

Deep Search

Deep Search is an agentic code search tool that understands natural language questions about your codebase. Unlike traditional search, Deep Search uses AI to intelligently explore your code, following leads and synthesizing information to provide comprehensive, conversational answers.

• Ask questions in natural language and get detailed, sourced answers about architecture, implementations, and code patterns
• Use @-mentions to scope searches to specific repositories, files, directories, or symbols
• Follow up with additional questions to drill deeper into specific aspects of your codebase
• View all sources — see exactly which searches were performed and which files were analyzed

​

Code Navigation

Sourcegraph gives your development team cross-repository, IDE-like features directly in the browser—no local checkout required:

• Go to definition — jump to where a symbol is defined, even across repositories
• Find references — enumerate all call sites and usages across your entire codebase
• Hover tooltips — inline documentation and type information
• Symbol search — navigate directly to functions, types, and variables by name
• Find implementations — navigate from interface definitions to all implementing types

Sourcegraph offers both search-based code navigation (available for 40+ languages with no setup) and precise code navigation using SCIP/LSIF data, which delivers compiler-accurate results with no false positives across 11 languages including Go, TypeScript, Java, Rust, and Python.

​

Batch Changes

Automate large-scale code changes across many repositories and code hosts. Batch Changes lets you write a specification for what to change, preview the results, and create pull requests on all affected repositories simultaneously. Common use cases:

• Update API callsites after a library upgrade
• Patch critical security vulnerabilities across all repos
• Rename symbols or update configuration files at scale
• Track migration progress until all PRs are merged

​

Code Insights

Reveal high-level information about your codebase at its current state and over time. Code Insights turns any Sourcegraph search query into a visualization you can track on customizable dashboards. Use Code Insights to track:

• Migration progress (e.g., framework or library upgrades)
• Version adoption across all repositories
• Vulnerability remediation status
• Codebase size trends and code smell patterns
• Ownership and component deprecation

​

Deployment options

Sourcegraph is available in two deployment models:

​

Who uses Sourcegraph?

Sourcegraph is a good fit for your team if:

• You have a large number of repositories or a large, complex monorepo
• You host code across multiple code hosts (GitHub, GitLab, Bitbucket, Perforce, or others)
• You rely heavily on “go to definition” and “find references” in your editor and want that capability at the codebase level
• You need to make coordinated changes across many repositories
• You want visibility into codebase-level trends over time

​

Integrations

Sourcegraph integrates with your existing tools:

• Code hosts — GitHub, GitLab, Bitbucket Cloud, Bitbucket Server/Data Center, Perforce, and other Git hosts
• Editors — VS Code, JetBrains IDEs, Gitpod
• Browsers — Chrome, Firefox, and Safari extensions add code navigation directly on GitHub, GitLab, and Bitbucket
• MCP Server — connect AI tools and IDEs to Sourcegraph’s code intelligence via the Model Context Protocol

​

APIs