Sectr
Sectr deploys agents built with LangGraph, OpenAI Agents SDK, or Pydantic AI, recording their conversations and handling approval waits. Agent developers use it to connect agents to frontends and resume them after human decisions.
Who it's for
- Developers with an existing LangGraph agent who want deployment and transcript-backed resumption without adding a checkpointer.
- Teams building agents with OpenAI Agents SDK or Pydantic AI that need human approval for tool calls.
- Frontend developers using CopilotKit who want to connect to an agent through an AG-UI endpoint.
Not the right fit if…
- Teams using an agent framework outside LangGraph, OpenAI Agents SDK, or Pydantic AI: the page documents adapters and templates for those three frameworks and does not specify support for others.
- Teams evaluating post-beta costs before adopting: the page says the product is free during beta but does not list future pricing.
How it fits your workflow
Initialize a project from a Sectr template, add the adapter and route to an existing agent, and use `sectr dev` to test it in a local chat UI.
Run `sectr deploy` to build and deploy the agent. Sectr provides an AG-UI endpoint, while the transcript records messages, tool activity, approvals, and errors.
When a tool requires approval, Sectr suspends the agent while waiting and resumes it after a decision. Developers can inspect and replay sessions in the Sectr dashboard.
Pricing
Free: Sectr is free during beta. The page does not specify a separate free-tier allowance or post-beta pricing.
Beta
Free during beta
- · Sectr agent deployment and platform access during beta
Prices checked on Oct 29, 2026 from the vendor's site. They can change; confirm before you buy.
Key features
- Framework adapters
- Add Sectr to agents built with LangGraph, OpenAI Agents SDK, or Pydantic AI using the corresponding adapter.
- Agent transcript and state
- Sectr records messages, tool calls and results, approval decisions, and errors in order. The transcript supplies conversation history and is used to rebuild framework state when an agent resumes.
- Human approval suspension and resumption
- Agents can pause when a tool requires approval, scale to zero while waiting, and continue after a person responds.
- AG-UI endpoint and streaming
- Deployed agents receive an AG-UI endpoint. The site describes streaming text, tool calls, and approvals to a frontend, including CopilotKit.
- Local development chat UI
- `sectr dev` lets developers test an agent locally in a chat UI without first writing frontend or server code.
- Cloud deployment
- `sectr deploy` builds and deploys an agent in the cloud and provides an AG-UI endpoint.
Works with
- LangGraph
- OpenAI Agents SDK
- Pydantic AI
- AG-UI
- CopilotKit
Limitations to know
- The site says Sectr is free during beta but does not provide pricing for after the beta.
- The page documents adapters and templates for LangGraph, OpenAI Agents SDK, and Pydantic AI; it does not specify support for other agent frameworks.
Alternatives to consider
-
Superserve
Choose it if you need a self-hostable sandbox with durable Firecracker microVMs and SDK/API control for long-running agents; Sectr focuses on framework adapters, transcript-backed state, approvals, and AG-UI.
Getting started
sectr init my-agent --template openai-agents- Sign up with GitHub, Google, Vercel, or email.
- Initialize a project using a template for a supported framework and add the Sectr adapter to the agent.
- Run `sectr dev` to try the agent in the local chat UI.
- Run `sectr deploy` to deploy the agent and get an AG-UI endpoint.