Integrations

Every way to get Weckr in front of your LLM calls. Wrap the provider client directly for the cleanest, lowest-overhead path, or use one of the framework guides for agent-style code.

No matter which path you pick, the dashboard at app.useweckr.com fills with the same per-user, per-feature cost data.

SDKs

TypeScript / JavaScript

→

The @weckr/sdk package for Node, Bun, Deno, and edge runtimes.

Python

→

weckr-sdk on PyPI. Works with sync OpenAI, Anthropic, and Gemini clients.

LLM providers

OpenAI

→

GPT-4o, GPT-4o-mini, GPT-4, GPT-3.5, o1-preview, o1-mini.

Anthropic

→

Claude Opus 4, Sonnet 4, Haiku 4.5, plus the 3.5 family.

Google Gemini

→

Gemini 2.5 Pro and Flash. Works with both @google/genai and the legacy SDK.

Agent frameworks

LangChain

→

Use Weckr alongside LangChain by wrapping the underlying OpenAI client at the boundary.

CrewAI

→

Catch runaway agent loops with token-velocity alerts. Same boundary-wrap pattern as LangChain.

LlamaIndex

Coming soon

RAG pipelines and query engines. Native callback handler on the roadmap.

Tooling

MCP server

→

Talk to your Weckr project from Claude Desktop, Cursor, or any MCP host.

Don't see your framework?

Tell us what you're using on GitHub and we'll prioritise a guide. Until then, the boundary-wrap pattern (wrap the underlying provider client with wk.chat()) works with almost any agent framework.