Terminal Use
Terminal Use is a managed infrastructure platform for deploying and scaling background AI agents with filesystem access, git-native branching, and built-in observability.
At a Glance
About Terminal Use
Terminal Use is a purpose-built infrastructure platform for deploying Claude Agent SDK and Codex agents in production. It provides managed compute environments with persistent filesystem storage, git-native versioning, and real-time streaming — so developers can focus on agent logic rather than infrastructure. Backed by Y Combinator, Terminal Use handles orchestration, execution, and delivery for filesystem-native AI agents across code generation, document processing, security audits, and more.
- Managed Agent Infrastructure: Deploy agents as Docker images with dedicated VMs, no cold starts, and no time limits — just run
npx skills add terminaluse/agent-skillsandtu deployto get started. - Git-Native Versioning & Rollback: Every deployment is versioned with instant rollback, blue-green deployments, and git metadata tracking to safely manage production agents.
- Persistent Filesystem per Task: Each task gets isolated, persistent storage; mount shared resources read-only or per-user workspaces read-write for true multi-tenant isolation.
- Multi-Environment Support: Run prod, staging, and preview environments with per-environment secrets, branch-based routing, and automatic environment selection.
- Built-in Observability: Automatic log collection, task history & replay, performance metrics, and real-time log streaming without external tooling.
- Flexible Triggers: Start agent tasks via API/SDK, CLI, webhooks, or scheduled runs from GitHub, Linear, Slack, or any webhook source.
- Vercel AI SDK Integration: Stream agent responses directly to React frontends using the native
@terminaluse/vercel-ai-sdk-providerpackage. - Python & TypeScript SDKs: Install with
pip install terminaluseor the npm package; scaffold, deploy, and manage agents entirely from the terminal or via your coding agent. - Enterprise Security: Fine-grained RBAC, encrypted secrets at rest, workload identity, and isolated data and compute between customers.
- Task Forking: Fork from any checkpoint to test fixes in dev and ship corrections to prod while preserving filesystem snapshots and conversation history.
Community Discussions
Be the first to start a conversation about Terminal Use
Share your experience with Terminal Use, ask questions, or help others learn from your insights.
Pricing
Hobby
Free tier for individual developers to start building agents.
- 1 agent
- 1 GB filesystem storage
- 100 tasks / month
- Community support
Starter
For small teams and growing projects.
- 3 agents
- 10 GB filesystem storage
- 1,000 tasks / month
- Community support
Pro
For production workloads with unlimited tasks and priority support.
- 10 agents
- 50 GB filesystem storage
- Unlimited tasks
- Priority support
Enterprise
Custom plans for larger teams — schedule a call or email founders.
- Custom agent limits
- Custom storage
- Dedicated support
- Enterprise security and RBAC
Capabilities
Key Features
- Managed agent infrastructure
- Git-native versioning and rollback
- Persistent filesystem storage per task
- Multi-environment support (prod, staging, preview)
- Built-in observability and log streaming
- API/SDK, CLI, webhook, and scheduled triggers
- Vercel AI SDK integration
- Python and TypeScript SDKs
- Fine-grained RBAC and encrypted secrets
- Task forking from checkpoints
- Docker image support
- Real-time message streaming
- Blue-green deployments
- Workload identity and multi-tenant isolation
Integrations
Demo Video

