Build once. Route every prompt.
Laminarity provides OpenAI-compatible Chat Completions and Models APIs, plus a stateless Responses beta. With model: "auto", it selects an eligible model from your configured provider pool for each request.
Route automatically, or select a model explicitly
GET /v1/models now returns the models allowed by the authenticated key's project policy. Use model: "auto" for Laminarity selection and fallback. A concrete model id makes a direct, no-fallback request.
Core API
What changed
Broader SDK compatibility
Chat Completions and Responses, plus developer messages, completion-token controls, strict tools, structured output fields, and more.
Predictable errors
Validation failures use an OpenAI-shaped 400 error. Unknown concrete models return 404 with code: "model_not_found".
Forward-compatible requests
Unknown request fields are ignored, reducing breakage as OpenAI-compatible SDKs add optional fields.
Portable feature set
Provider capabilities differ. For auto workflows, prefer the features Laminarity translates across the routing pool.
| Portable across providers | Provider-dependent |
|---|---|
| Messages, streaming, tools, vision input | Strict tool schemas and parallel-tool controls |
| Temperature, top-p, stop sequences | Structured output implementation details |
max_completion_tokens | Reasoning, service-tier, caching, and prediction fields |
Integration guides
Flow adaptive routing
Train an adaptive router from your project's model choices.
Read guideQuickstart
Make an authenticated request, understand auto, and inspect the selected provider.
Responses API (beta)
Use Laminarity's stateless OpenAI Responses-compatible subset for response, tool and coding-agent flows, including OpenAI Codex CLI.
Read guideAny OpenAI-compatible app
Connect Laminarity to a new or unfamiliar application that accepts a custom OpenAI-compatible provider.
Read guideHermes Agent
Connect Hermes Agent through a named OpenAI-compatible provider while preserving Laminarity routing.
Read guideOpenAI SDKs
Use the official Python or TypeScript SDK with Laminarity's base URL.
Read guideVercel AI SDK
Stream a server-side chat through the OpenAI-compatible provider adapter.
Read guideLangChain
Point ChatOpenAI at Laminarity for invoke, stream, and tool-enabled agent flows.
Read guideOpenCode
Configure Laminarity as an OpenAI-compatible OpenCode provider with model discovery, tools, and explicit reasoning variants.
Read guideLangChain4j
Use Laminarity from plain Java, streaming Java, Spring Boot, or a LangChain4j-backed agent provider.
Read guideKilo Code
Add Laminarity as a custom OpenAI-compatible Kilo Code provider with model discovery and explicit capabilities.
Read guideCline
Connect Cline to Laminarity through its OpenAI-compatible provider settings.
Read guideOpenClaw
Configure an explicit OpenAI Completions provider for Laminarity without enabling native-provider request shaping.
Read guideAider
Point Aider's OpenAI-compatible client at Laminarity for routed or model-pinned coding sessions.
Read guideContinue
Configure Continue's IDE or CLI agent to use Laminarity through Chat Completions.
Read guideLlamaIndex
Use LlamaIndex's OpenAILike adapter for Laminarity chat and streaming workflows.
Read guideOpen WebUI
Connect Open WebUI to Laminarity with authenticated model discovery and Chat Completions.
Read guideLiteLLM
Call Laminarity through LiteLLM's OpenAI-compatible client or expose it behind a LiteLLM proxy alias.
Read guide