Sourcegraph Cody
Sourcegraph Cody is an AI coding assistant developed by Sourcegraph that provides code autocomplete, explanation, and refactoring across your entire codebase. It differentiates itself by understanding your full repository context for more accurate, c
Sourcegraph Cody is an AI coding assistant that uses Sourcegraph's code-search engine to pull deep context from large, multi-repo codebases for chat, autocomplete, inline edits and code explanation inside the IDE. As of mid-2026 it has been repositioned as an enterprise-only/legacy product: Sourcegraph discontinued the Cody Free and Pro tiers on July 23, 2025, and now steers individual developers to Amp instead, while continuing to support Cody for large-org Enterprise customers. It remains notable mainly for codebase-scale context retrieval rather than autonomous agentic coding.
- Built by
- Sourcegraph
- Category
- IDE extension / AI coding assistant (enterprise, codebase-context-focused; not an autonomous agent)
- Open source
- Yes
- Models
- Multi-model and model-agnostic. Supports Anthropic Claude (Opus, Sonnet, Haiku 4.x family, including thinking variants), OpenAI GPT-5 series and GPT-4o plus o3/o4-mini reasoning models, and Google Gemini (2.5 and 3 variants). Default chat model is Claude Sonnet 4.5; default autocomplete model is DeepSeek V2 Lite Base. Enterprise admins can select/enforce models and bring their own keys via Amazon Bedrock, Google Vertex AI, or Azure OpenAI.
Available on
Pricing
Enterprise-only as of 2026. Cody Enterprise is roughly $59/user/month on an annual contract; the broader Sourcegraph Enterprise platform (Code Search + AI features) is quoted starting around $16K and scales with team size. There is no longer a free or self-serve Pro tier — Cody Free and Cody Pro were discontinued on July 23, 2025 (new signups stopped June 25, 2025), and individual users are directed to Sourcegraph's separate product Amp.
Key capabilities
- ▸Codebase-wide context retrieval powered by Sourcegraph's code-search engine across local and remote/multi-repo codebases
- ▸Multi-model flexibility — admins choose and enforce Claude, GPT, or Gemini, including bring-your-own-key (Bedrock, Vertex AI, Azure OpenAI)
- ▸Chat, autocomplete, inline edit, and explain-code across VS Code, JetBrains, Visual Studio and Web
- ▸Enterprise security and compliance: SOC 2, no training on customer code, self-hosted / air-gapped deployment options
- ▸Admin controls for model policies, access, and team-wide configuration
- ▸Integration with the wider Sourcegraph platform (Code Search, Deep Search, MCP server, API, CLI)
Strengths
- +Best-in-class context over very large, multi-repo enterprise codebases via Sourcegraph search
- +Model-agnostic — not locked to one LLM vendor, with BYO-key and self-hosting for compliance
- +Strong enterprise security/governance posture (SOC 2, no-training guarantee, air-gapped deploys)
- +Open source (Apache 2.0) client, giving transparency and extensibility
Limitations
- −No longer available to individuals or small teams — Free and Pro tiers were discontinued in July 2025, leaving only a costly enterprise plan (~$59/user/month plus platform cost)
- −Effectively a legacy/maintenance product: Sourcegraph's strategic focus and its founders moved to the separate Amp product/company, so Cody lags newer agentic tools
- −More an assistant than an autonomous coding agent — weaker multi-step task execution than agentic competitors, and no published agent benchmark scores (e.g. SWE-bench) for Cody itself
Best for
Large enterprise engineering orgs with huge multi-repo codebases, existing Sourcegraph adoption, and strict security/compliance needs — not individual developers (who are pointed to Amp).
What's new
In December 2025 Sourcegraph split into two independent companies: co-founders Quinn Slack and Beyang Liu left to found Amp Inc. (taking the Amp agentic-coding product), while Sourcegraph under new CEO Dan Adler retained Code Search and continues to support Cody Enterprise as a legacy product. Cody Enterprise has continued receiving new models (e.g. Claude Opus 4.5, GPT-5.1, Gemini 3 Pro) across VS Code, JetBrains, and Cody Web.
Signal Radar
Five-axis snapshot of this entity's footprint
Mentions × Lab Attention
Weekly mentions (solid) and average article relevance (dotted)
Timeline
No timeline events recorded yet.
Relationships
2Frequently appears with
1Entities that show up in the same articles — shared coverage, not a stated relationship.
Recent Articles
No articles found for this entity.
Predictions
No predictions linked to this entity.
AI Discoveries
No AI agent discoveries for this entity.
Sentiment History
| Week | Avg Sentiment | Mentions |
|---|---|---|
| 2026-W23 | 0.05 | 2 |