New AI agent enables engineering teams to plan, execute, and manage complex code changes across hundreds or thousands of repositories while keeping engineers in control.

Sourcegraph, the code intelligence platform that helps enterprise engineering teams understand, oversee, and evolve their codebases, announced public beta availability of Agentic Batch Changes. Agentic Batch Changes is an AI agent that builds on Sourcegraph’s Batch Changes execution engine and code search. By bringing AI-powered planning and orchestration to large scale code changes, it enables teams to plan, execute, and track changes across hundreds or thousands of repositories, starting with a single prompt.

The problem with large-scale code change

Every engineering team eventually faces the same challenge: a change that needs to happen everywhere. A dependency needs upgrading. An API changes and every caller needs updating. A security vulnerability drops and the fix needs to reach every exposed repository before the next audit window.

The hard part is not deciding what change to make. It’s executing that change across hundreds of repositories without losing visibility, introducing inconsistency, or pulling engineers into weeks of repetitive manual work.

Until now, engineering teams had four options. Make the changes manually. Write brittle one-off scripts. Ask a coding agent to clone hundreds of repositories locally and trust it to make it through the rollout. Or avoid the rollout entirely and let the backlog grow.

A new approach

With Agentic Batch Changes, teams start with a single prompt describing the change they want to make. The agent uses Sourcegraph search and Deep Search to identify repositories that need the change, validates its approach in an initial repository, then expands the rollout. It handles repository specific variation, reacts to CI signals, and iterates on failures before publishing at scale. Engineers review and approve every changeset before merge.

The result is faster organization-wide code change executed with a human in the loop.

”Agentic Batch Changes brings the capabilities of the best coding agents to the largest companies in the world,” said Dan Adler, CEO of Sourcegraph. “For years, the owners of the largest codebases have watched smaller competitors move faster because large scale changes were too slow and too risky. Agentic Batch Changes changes that, enabling engineering teams to roll out changes across thousands of repositories and directories with the speed and confidence of changing just one.”

Mercari used Agentic Batch Changes to scope a security vulnerability

Mercari, the Japan-based marketplace platform, has been using Agentic Batch Changes during the experimental preview. Patrick Klitzke, Team Lead at Mercari, used Agentic Batch Changes to identify a GitHub Actions environment-variable injection vulnerability across Mercari’s codebase and began patching it across the affected repositories.

“I looked at a GitHub injection issue where you have to set environment variables correctly. I was able to fix it with one prompt on both the Help Center frontend and backend, then extended this to all repos in Mercari – I found around 80 potential repos affected,” Klitzke said.

Klitzke noted that the agent’s ability to reason about each repository individually is what made the workflow possible.

“With the help of Agentic Batch Changes, you’re able to handle repos that have similar, but not identical setups. A normal scripted change would most likely be a text search and replace operation without any context of how it’s actually used,” Klitzke said.

What Agentic Batch Changes can do

Agentic Batch Changes is designed for code changes that have historically been too complex to script, including:

• Dependency upgrades with breaking changes. Major-version bumps where the API surface shifts. The agent handles per-repo variation.
• CVE remediation. Sourcegraph search identifies every instance of a vulnerable library. Agentic Batch Changes ships the fix across every affected repository.
• Code pattern updates with judgment. Refactors that need context, not regex. Internal style migrations, deprecated API replacement.
• New API or library rollouts with per-service variation. New auth contracts, new third-party libraries, service mesh changes.
• CI pipeline modernization. Build-system changes, test infrastructure overhauls.

The best approach for every code change

Built on Sourcegraph’s code graph, Deep Search, and Batch Changes execution infrastructure, Agentic Batch Changes chooses the right execution strategy for every change. For mechanical changes, it generates deterministic scripts or an entire program to apply updates at scale. For changes that need judgment, it delegates to frontier coding agents such as Claude Code and Codex, providing repository specific instructions codebase context through the Sourcegraph MCP.

Availability and pricing

Agentic Batch Changes is available in public beta starting June 30, 2026, for Sourcegraph Cloud customers at no additional cost during the beta period. Self-hosted availability begins with Sourcegraph 7.5 on July 8, 2026.

Source: BUSINESS WIRE