A new era for Sourcegraph: The intelligence layer for AI coding agents and developers

Today, we're announcing Sourcegraph 7.0, a release that marks the beginning of a new chapter for our company and product.

Over the past several releases, we've shifted our focus. We're doubling down on being the intelligence layer that developers and AI agents rely on to navigate, understand, and operate on large codebases.

This release formalizes that shift.

The challenge: Agent context at enterprise scale

Teams everywhere are experimenting with AI coding agents. The promise is real: faster development, better code, more time spent on creative problem-solving instead of mechanical work.

But there's a gap. AI agents struggle with the same challenges developers face when working with large, complex codebases: understanding cross-repository dependencies, reasoning about historical context, and navigating architectural decisions made years ago.

The result is shallow context, misleading suggestions, and a lack of trust in agent output when it matters most.

Sourcegraph as shared intelligence

Sourcegraph 7.0 addresses this by positioning our platform as the shared intelligence layer for both developers and AI agents.

This isn't about a single feature. It's about a fundamental shift in how we think about code intelligence. Humans and agents are both first-class users, working through the same underlying system to reason about code and safely make changes at scale.

When an AI agent needs to understand your codebase, it should have access to the same deep, accurate context that your best developers rely on. When a developer needs to make a change across multiple repositories, they should benefit from the same semantic understanding that powers agent workflows.

What's new in 7.0

Deep Search via MCP

AI agents can now ask semantic, cross-repository, historical, and architectural questions about your entire enterprise codebase through our MCP server. This means agents can:

Understand how systems connect across repository boundaries Reason about why code was written the way it was Navigate complex architectural patterns with confidence Access the same reliable context developers trust

This closes the biggest gap in agentic development workflows: dependable, shared code context. Teams will also be able to track MCP tool usage in Sourcegraph Analytics.

Better-than-ever Deep Search, integrated in your workflows

We've improved Deep Search to deliver even more accurate results, added image support, and made it available wherever you need it. Whether you're a human developer investigating a bug or an AI agent planning a refactoring, the same powerful semantic understanding is available.

For those customers looking to create custom integrations we're rolling out a new versioned Sourcegraph API. We love integrations so much we've now integrated Code Navigation right into Deep Search, now when you get results return from your search you can explore them right there in the sidebar.

Foundation improvements

Beyond the headline capabilities, we've strengthened the foundation:

Modernized Svelte UI for a faster, more responsive experience Improved instance setup flow to get teams up and running quickly Weekly shipping cadence for Sourcegraph Cloud so improvements reach you faster

7.0 feature removals

Sourcegraph has always made a commitment to ship fast and with confidence. Moving fast means doubling down on things that are working while making changes to directions we've explored in the past. As such, we're removing and deprecating some features, including Cody Web for Deep Search customers and Search Notebooks, along with a few more that are worth checking out in our changelog update.

What we're not promising

We care deeply about trust, so it's important to be clear about what this release is and isn't.

We're not claiming that agents write perfect code. We're not claiming that Sourcegraph replaces human judgment. And we're not positioning this as a flashy UX overhaul.

This is about infrastructure correctness. It's about giving both developers and AI agents the reliable, deep context they need to work effectively with real production code. The rest is up to you.

What this means for you

If you're already a Sourcegraph customer, 7.0 unlocks immediate new value from the code you already have indexed and the workflows you already run. Your investment in Sourcegraph becomes foundational infrastructure for the AI-assisted development era, not just a search tool.

If you're evaluating how to enable effective agent workflows at enterprise scale, Sourcegraph 7.0 provides the authoritative intelligence layer that both your developers and your AI agents can depend on.

Looking ahead

Development is becoming increasingly AI-assisted. The question isn't whether agents will be part of your workflow, but whether they'll have the context they need to be helpful rather than harmful.

Sourcegraph 7.0 is our answer: a shared intelligence layer that helps teams work confidently with complex code, whether they're human developers or AI agents.

This is the beginning of a new era for Sourcegraph. We're excited to build it with you.