Grok Build
Grok Build is xAI’s terminal coding agent for repository work. Official documentation supports three primary modes: an interactive full-screen terminal interface, headless execution for scripts or bots, and Agent Client Protocol integration inside other applications. It can inspect a codebase, plan a task, edit files, run development commands, and present work for review. The product page also documents parallel subagents, Git worktrees, plugins, reusable skills, hooks, and Model Context Protocol servers for extending its workflow.
The current xAI page provides a shell installer. First launch normally opens a browser for authentication; non-browser environments can use an xAI API key. Official docs also allow configuration of an OpenAI-compatible custom model endpoint. xAI’s current product page says Grok Build is available to try for free, but it does not publish a durable entitlement table. xAI publishes the Grok Build CLI and TUI source under Apache License 2.0 at xai-org/grok-build; that public agent-runtime source is not the same thing as downloadable Grok model weights.
Top features
- Mouse-capable, full-screen terminal UI for interactive coding sessions.
- Plan mode that blocks edits until the user reviews and approves the proposed approach.
- Headless prompt mode with structured output options for scripts, bots, and integrations.
- Agent Client Protocol support for embedding the agent in compatible clients.
- Parallel subagents for research, implementation, testing, and review.
- Git worktree support for isolating concurrent subagent work.
- Plugins and marketplaces that can bundle skills, agents, hooks, and MCP servers.
- Custom model configuration through an OpenAI-compatible endpoint, alongside xAI authentication.
Use cases
- Inspect a repository and ask for an architecture explanation before making a change.
- Plan and implement a multi-file refactor with an explicit approval checkpoint before edits.
- Run separate subagents for investigation, implementation, and testing in isolated worktrees.
- Invoke a bounded code-analysis or documentation task from a script using headless output.
- Connect approved engineering tools such as issue trackers, error monitoring, or databases through MCP.
- Add a coding agent to an ACP-compatible application without rebuilding the terminal workflow from scratch.
No linked agents yet.
Register an agentSpaceXAI Launches Grok 4.6 With a Focus on Long-Running Agents
SpaceXAI has released Grok 4.6, a model aimed at long-running agents, coding, knowledge work, and interactive projects. It is available in Cursor, Grok Build, the SpaceXAI API, and selected model gateways at the same base token price as Grok 4.5.
AI Agents · Aug 12, 2026
Grok Build 1.0 Is xAI’s Bid to Make Terminal Coding Agents Usable Every Day
Grok Build 1.0.0 arrived on August 7, 2026. Its headline is not a new model or benchmark score, but the less glamorous work of making an agent that edits code, runs commands, and keeps working easier to control and recover.
AI Tools · Aug 8, 2026
Cloudflare Built a Browser for Agents, Not Humans, and Put It on Workers
Kitesurf is a Chromium alternative for agent workloads: Rust/Wasm pieces, V8 isolates, CDP compatibility, 3–7× less CPU/memory on Cloudflare’s corpus, and an honest list of things it still cannot do.
ai-agents · Aug 7, 2026

