Open this AI tool now
https://sourcegraph.com/codyIntroduction
Sourcegraph Cody provides engineers with a practical capability to turn code repositories into a queryable knowledge source with instant editing, making exploration, editing, and documentation across large codebases faster and more accurate than relying on manual search or individual memory.
What is the tool?
Sourcegraph Cody is an AI‑powered paid programming assistant designed to work with source code and repositories. Cody combines Sourcegraph’s advanced code search techniques (such as structural search and index‑based context) with Retrieval‑Augmented Generation (RAG) mechanisms that rely on context embeddings, along with large language model (LLM) engines for generation, explanation, and completion.
Tired of juggling ten tabs? ToolSuite bundles the AI workflow tools power users rely on — in one place.
Try ToolSuite NowCody integrates with:
- Sourcegraph’s web interface (Cody Chat in the browser).
- Popular IDE extensions such as VS Code and JetBrains.
- Multiple hosting approaches: Sourcegraph Cloud or self‑hosted (Cody Server) for privacy and security requirements.
Functionally, Cody provides contextual chat about code (Cody Chat), code suggestions and completion based on repository context, review‑ and pull request (PR)‑specific features, generation of proposed changes (diffs), explanation of functions and comments, and means to write tests and runbooks based on the indexed project tree.
Key Features
Repository‑aware context retrieval (Repository‑aware RAG)
Before generating the answer, Cody searches within the Sourcegraph index for the most relevant code and document fragments using embeddings and precise match signals, then provides these snippets to the model to respond. This reduces reliance on the model’s limited context window and increases source traceability.
Integrated Contextual Chat (Cody Chat)
A chat interface that allows asking questions about functions, dependencies, or test errors. The responses display reference snippets with links to locations in the repository, and it supports commands such as “give me the diff for changing X” or “create a PR with this change”.
Suggestions and code completion inside the IDE
Extensions work to provide suggestions based on multiple files in the project, not only on the currently open file; which makes completion aware of the project’s public interfaces (APIs) and architectural structures.
Precise support for code intelligence (Precise code intelligence)
When setting up LSIF or Sourcegraph indexing, Cody can link symbols across files and provide accurate answers such as definition and usage locations, which is helpful when explaining dependencies or tracing the source of an error.
Automated activities to create or modify PRs
The user can request “create a PR that fixes X” and Cody will suggest a set of changes (diff) that can be reviewed locally or via the interface to be pushed as a branch and a pull request.
Flexibility in providers and models
Supports running Cody with external LLM providers (such as OpenAI or internally managed models) or a self‑hosted architecture via Cody Server, to meet confidentiality and compliance requirements.
Security policy for private content (Private code handling)
Options to run indexing within the internal network or on the Cloud with access control, with the ability to retain a log of queries and policies to control what is sent to external LLM providers.
How to Use
Prerequisites
- A configured Sourcegraph Cloud account or Sourcegraph Self‑hosted instance.
- Access to repositories (GitHub/GitLab/Bitbucket) and linking them to Sourcegraph.
- Optional: An API key for a language model provider.
Comments
0No comments yet.
Please log in to comment.
Comments are available for members only. Sign in to participate in the discussion, or create a new account for free.