An individual developer can no longer sign up for Sourcegraph Cody. Sourcegraph stopped taking new Cody Free and Cody Pro sign-ups on 25 June 2025 and discontinued both plans on 23 July 2025, naming Amp, its newer tool built for agentic workflows, as the path forward for those users; Cody Enterprise was not affected. As of 4 October 2026, Sourcegraph’s pricing page shows one plan, Enterprise, from $16K as a minimum annual contract, with no pound price published. The sections below on context, features and the comparison with Copilot and Cursor describe what Cody is built to do, and they now matter to teams buying Sourcegraph, not to someone choosing an editor extension for personal use.
This review is expert analysis drawn from Sourcegraph’s documentation, current pricing, and aggregated developer feedback through mid-2026 — not a private benchmark or a claim of hands-on lab testing. What changed (4 October 2026): the earlier version of this review listed a free tier and a Pro plan at around $9 / £8 a month. Sourcegraph’s own announcement (Sourcegraph blog, Kelvin Yap, 25 June 2025) shows both plans ended on 23 July 2025, so that advice was wrong and has been corrected. Pricing and availability were re-read on Sourcegraph’s pricing page and Cody documentation on 4 October 2026. The feature and comparison sections have not been re-assessed.
Cody at a glance
Pros:
- Best-in-class codebase context, powered by Sourcegraph code search
- Excellent for large monorepos and unfamiliar, sprawling projects
- Model choice across leading providers, not locked to one
- Strong enterprise story: security, admin controls, code intelligence
- Works in VS Code, JetBrains and more
Cons:
- Full power needs the Sourcegraph platform, especially at enterprise scale
- Inline completion is good but not clearly ahead of rivals
- More setup and concept overhead than a plug-in-and-go assistant
- No plan for individuals: Cody Free and Cody Pro ended in July 2025
- Overkill for small solo projects
Cody’s core idea: search-grounded context
Every AI coding tool talks about “context.” Cody’s advantage is that context is its heritage. Sourcegraph spent years building a code-search and code-intelligence platform used to navigate huge codebases, and Cody plugs the AI layer directly into that engine. The practical result is that when you ask Cody a question or ask it to write something, it can retrieve the genuinely relevant files, symbols and usages from across the whole repository — not just guess from what is on screen.
That matters most exactly where AI assistants usually fail: large, interconnected codebases where the answer depends on code you did not know existed. Ask “how is authentication handled here?” in a sprawling backend and Cody can pull the real implementation from across services and explain it, rather than hallucinating a plausible-but-wrong pattern. For onboarding onto an unfamiliar project or working in a monorepo, this is a distinct, defensible edge. Cody is one of the standouts in our Best AI Coding Assistants 2026: Top Picks Compared & Ranked guide precisely for this reason.
Features and the editor experience
Cody delivers the expected assistant toolkit — inline autocomplete, a chat panel, edit/fix commands, test and docstring generation — inside popular editors including VS Code and the JetBrains IDEs. In day-to-day use the chat is the star: because answers are grounded in retrieved context, they are more trustworthy on real questions about your code than assistants relying on a narrow window.
Autocomplete is solid and useful, but honestly it is not where Cody pulls ahead. Several rivals offer completion that feels just as quick, and some in-editor experiences feel more seamless line to line. If low-latency single-line completion is the only thing you care about, Cody is not a clear winner. Its value concentrates in the harder, context-hungry tasks: understanding, refactoring and reasoning across a big codebase.
A genuinely useful trait is model flexibility. Cody lets you pick among leading models from different providers rather than tying you to one, so you can favour speed, quality or cost per task. That openness is increasingly valuable as the model landscape shifts — a theme we track across our The AI Directory.
Enterprise and team fit
Cody’s natural home is the organisation, not the hobbyist. On the enterprise tier it connects to your Sourcegraph deployment, bringing admin controls, security and compliance posture, and code intelligence across all your repositories. For a company trying to make a large, multi-team codebase legible — and to give every developer AI grounded in that shared context — the combined Sourcegraph-plus-Cody platform is compelling in a way single-file assistants are not.
That strength comes with a trade-off: to get the fullest experience, you are adopting a platform, not just installing an extension. Teams already using Sourcegraph gain the most and the setup feels natural; teams starting cold should weigh the additional concepts and configuration. For a solo developer on modest projects, that overhead is simply more than the payoff.
Pricing
Free and Pro: discontinued. Sourcegraph closed new sign-ups for Cody Free and Cody Pro on 25 June 2025 and ended both plans on 23 July 2025. Its announcement named Amp as the path forward for people on those plans.
Enterprise: Sourcegraph’s pricing page, read on 4 October 2026, lists one plan: Enterprise, starting at $16K as a minimum annual contract, with credits for AI features included and pricing that scales with team size. The page does not mention Cody by name; Sourcegraph’s Cody documentation describes it as supported on Sourcegraph Enterprise and sends interested teams to its sales contact.
US vs UK: the only published figure is in US dollars. Sourcegraph lists no pound price, so there is no UK list price to compare it with.
The honest read on value: there is no longer an individual tier to weigh against cheaper rivals like Copilot. The money argument was always strongest at team and enterprise scale, where codebase-wide context and code intelligence solve problems that per-file assistants cannot, and that is now the only way to buy it. A five-figure annual minimum makes Cody a procurement decision, not a personal subscription.
Cody vs GitHub Copilot and Cursor
Against GitHub Copilot, the split is context depth versus polish and ubiquity. Copilot is superbly integrated, widely adopted and excellent at inline completion, and it now has a capable agent mode; see our GitHub Copilot Review 2026: Still the Default?. Cody counters with deeper, search-grounded whole-repo understanding. If your pain is “I need help typing the next lines,” Copilot is the smooth default. If your pain is “I need to understand and safely change a huge codebase,” Cody’s grounding is the differentiator.
Against Cursor, the comparison is platform philosophy. Cursor is an AI-first editor built around an agent and next-edit tab completion, brilliant for hands-on multi-file editing; our Cursor Review 2026: The AI Code Editor, Examined covers it. Cody is less about reinventing the editor and more about bringing rigorous code intelligence into the editor you already use — and about scaling that intelligence across an organisation.
Who should use Cody
Cody is the pick for developers and teams whose defining challenge is a large, complex codebase: monorepos, multi-service backends, and unfamiliar projects you must ramp up on fast. Enterprises already invested in Sourcegraph, and any team that values security-conscious, admin-controlled AI grounded in their real code, will get the most from it.
It is no longer available to solo developers, since the individual plans ended in July 2025, and it is a weak fit for any team wanting a zero-setup, install-and-forget tool. In those cases a lighter assistant delivers more of the value with less overhead.
Verdict
Sourcegraph Cody is the codebase-aware AI assistant, and in 2026 that framing is its whole strength. Built on a proven code-search engine, it grounds its answers and edits in how your actual repository works, which makes it uniquely good at large, interconnected codebases and at onboarding onto unfamiliar ones. Model choice and a serious enterprise story round out the package. The trade-offs are setup overhead, a completion experience that is good rather than category-leading, and diminishing returns on small solo projects.
Shortlist it if your team lives in a big monorepo and needs AI that understands the whole codebase, particularly if you already run Sourcegraph. If you are a solo developer, Cody is no longer on offer to you: the free and Pro plans ended in July 2025, and a lighter assistant such as Copilot or Cursor is the realistic choice.
FAQ
Is Sourcegraph Cody worth it in 2026?
For developers and teams working in large, complex codebases, yes. Cody’s search-grounded context — built on Sourcegraph’s code-search engine — makes it genuinely better at understanding and safely changing sprawling repositories than per-file assistants. The value is strongest at team and enterprise scale. For a solo developer on small projects who mainly wants fast inline autocomplete, the payoff is smaller and a lighter tool may suffice. There is no longer a free tier to try it on: Cody Free and Cody Pro ended on 23 July 2025, so access now comes through Sourcegraph Enterprise.
How is Cody different from GitHub Copilot?
The core difference is context depth. Copilot is beautifully integrated, ubiquitous and excellent at inline completion, with a capable agent mode. Cody’s edge is whole-repository awareness powered by Sourcegraph code search, so it grounds answers in your real codebase rather than just the open file. Choose Copilot for smooth day-to-day completion; choose Cody when understanding and changing a large, interconnected codebase is the actual challenge. See our GitHub Copilot Review 2026: Still the Default? for the full picture.
Does Cody support multiple AI models?
Yes. Cody lets you choose among leading models from different providers rather than locking you to one, so you can favour quality, speed or cost depending on the task. That flexibility is increasingly valuable as models improve at different rates, and it means you are not stranded on a single vendor’s roadmap. It is one of the features that makes Cody feel built for professional, long-term use.
What makes Cody good for large codebases?
Cody is built on Sourcegraph’s code-search and code-intelligence platform, so it can retrieve the genuinely relevant files, symbols and usages from across an entire repository when answering a question or writing code. That grounding is exactly what per-file assistants lack, and it is why Cody excels at monorepos, multi-service backends and onboarding onto unfamiliar projects where the answer depends on code you did not know existed.
How much does Sourcegraph Cody cost?
As of 4 October 2026 Cody has no free or Pro plan: Sourcegraph discontinued both on 23 July 2025. Its pricing page lists a single Enterprise plan starting at $16K as a minimum annual contract, with credits for AI features included, and Sourcegraph’s documentation describes Cody as supported on Sourcegraph Enterprise. No pound price is published.
Zen Tech Hub may earn a commission from links on this page, at no extra cost to you.
