Agentic coding tool from Sourcegraph for planning and implementing software changes with repository and tool context.
Amp is an AI coding agent from Sourcegraph that brings repository-aware task execution to software development. It can help developers move from a written requirement to concrete code changes, using tools and project context along the way. The product is suited to engineers who want an agent to make progress on bounded tasks instead of only returning an answer in a chat window.
Amp is Sourcegraph's agentic coding product, designed to handle development tasks with access to project context and developer tools. Instead of only suggesting a line of code, it can plan changes, edit files, and work through a task in a more agent-like loop. The tool is aimed at engineers who want delegated coding work while retaining control of the repository.
Agentic coding is most useful when the task can be clearly described and validated through tests or review. Amp should not be treated as a replacement for engineering judgement, particularly around architecture, security, or production operations. Its pricing and supported interfaces can change, so current plan documentation should be checked before adopting it for a team.
Developers who want a focused AI agent for repository-level coding tasks. A practical starting task is investigate a failure.
Developers who want a focused AI agent for repository-level coding tasks. A practical starting task is investigate a failure.
Delegates multi-step work
Keeps context near the code
Useful for implementation and refactoring
Paid
Amp is offered through paid usage plans, and current pricing is tied to model or plan consumption. Check the official Amp pricing information for current rates.
Developers who want a focused AI agent for repository-level coding tasks.
| Developer | Sourcegraph |
|---|---|
| Launched | 2025 |
| Platforms | Web, Windows, Macos, Api |
| Languages | English |
| API available | Yes |
| Open source | No |
