Codebases are growing and changing faster than ever, but the problem of maintaining systems - old and new - is not going away for codebase owners. If anything, it is becoming more important than ever. Today, we're taking the next step to meet this new era of software development with Agentic Batch Changes, available now in Beta: the frontier agent for code change at scale, built on technology designed for the world's largest codebases.
Agentic Batch Changes is built on top of Batch Changes, our precision tool for making coordinated updates across all your code. Paired with Deep Search, our agent for understanding complex codebases, these two tools form the foundation of an agent harness specialized for applying changes across codebases at any scale: from thousands of repositories to the largest monorepos.
This could mean anything from large-scale migrations, to patching security vulnerabilities, to applying upgrades with breaking changes, and more: anything a single coding agent can do in one repository, Agentic Batch Changes can orchestrate across all of your repositories at the same time.
The Agentic Batch Changes Beta is now available upon request: talk to your account team or schedule a demo to get started.
An agent for programming code change
Agentic Batch Changes works with you to prepare a plan. With a deep understanding of all your code through Deep Search, the agent can choose the right approach for the problem and target space, using a combination of strategies:
- For deterministic changes, the agent can opt to write a script - or even an entire program - to efficiently apply some, or all, of the target changes.
- For changes that require some additional judgement for each repository, the agent can decide to delegate to Claude Code or Codex to work on each target separately, with detailed instructions based on its plan. Coding agents are also integrated with the Sourcegraph MCP out of the box, giving them access to all the essential context.
The agent can then apply these plans using the batch change execution engine across all your repositories.
Building this agent, we learned that when you give a coordination agent a deep understanding of the problem space, and the tools to effectively produce and verify the change, there's a surprising amount of changes this agent can do without spending any inner-loop coding agent tokens at all or very few using a mix of determinism and highly specialized prompting, making this a highly effective harness for change at scale.
Just as your local coding agents benefit from verification tools (for example, tests and linters), so too you can ask Agentic Batch Changes to incorporate these tools into the plan. The human operator's role remains important as well: the agent will generate changes incrementally, reviewing them as it progresses, and providing opportunities for you to provide feedback on its plans.
But a diff is not a shipped change, so we're also introducing the ability to react to events that happen to changes when they get pushed to a code host, such as:
- When CI fails, the agent can prepare a script to fetch CI logs for you. This data can be used to amend the agent's plan, or even be handed off to a coding agent to fix immediately.
- When a merge conflict is encountered, the agent can ask a coding agent to address it.
These capabilities help the agent iterate on its changes until they're ready to be approved and merged.
Most organizations will also have specific conventions, technologies, and preferences. The agent may discover some of these conventions in its research, but you can also improve performance by customizing the agent with skills and AGENTS.md-style guidance, available on a per-user, per-organization, or global level in settings.
Get started today and help shape this tool
Sourcegraph is committed to solving the challenge of making the largest, most complex codebases in the world easier to understand, oversee, and evolve.
This launch of the Agentic Batch Changes Beta is the first piece to a much larger picture. We will continue to refine and evolve the product from here based on your feedback, and can't wait to take it to the next level.
Learn more at sourcegraph.com/agentic-batch-changes. Talk to your account team or schedule a demo to get started!