Outlines
by .txt
Guarantees LLM outputs conform to a schema, regex, or grammar through structured generation — reliable JSON, function calls, and constrained decoding. By .txt (dottxt-ai). 14k+ stars, Apache-2.0.
Skills
Constrained Decoding
Forces token generation to follow a regex, JSON schema, or context-free grammar so output is always valid.
Schema-Bound JSON
Generates JSON that provably matches a Pydantic or JSON Schema definition without any post-hoc parsing.
Grammar-Guided Output
Constrains model output to a formal grammar for structured formats like SQL or custom domain languages.
Related Agents
OpenSandbox
Runs AI-agent workloads in isolated Docker or Kubernetes sandboxes, exposing sandbox lifecycle, command, filesystem, an…
Browser Harness
Connects an LLM directly to a real Chrome browser over a single CDP websocket, letting the agent write the helper code…
Guidance
Programs LLM output with interleaved generation, control flow, and constraints — regex, grammars, and tool calls in one…
MCP Python SDK
Builds MCP servers and clients in Python with typed tool, resource, and prompt primitives plus stdio and streamable HTT…