@dawn-ai/langchain

LangChain backend adapters for Dawn, the TypeScript meta-framework for LangGraph that lets you build LangGraph agents like Next.js apps. Dawn uses this package to materialize chain routes and provider-aware agent routes — handling tool conversion, streaming, and retry.

agent() materialization resolves a LangChain chat model from the route descriptor. Dawn includes @langchain/openai for the default/backcompat path and lazy-loads optional provider packages when an agent selects or infers another provider.

OpenAI embedder

openaiEmbedder() provides Dawn's default semantic-memory embedder. Use it in dawn.config.ts to opt a memory-enabled route into hybrid keyword + vector recall:

import { config } from "@dawn-ai/core"
import { openaiEmbedder } from "@dawn-ai/langchain"

export default config({
  memory: {
    vector: { embedder: openaiEmbedder() },
  },
})

By default it uses text-embedding-3-small, reports dims: 1536, stores the embedder id as openai:text-embedding-3-small, and returns Float32Array vectors. Pass { model: "text-embedding-3-large" } for a 3072-dimension embedder.

openaiEmbedder() honors OPENAI_BASE_URL, the same seam used by Dawn's test harness and aimock. It explicitly requests OpenAI's float embedding encoding (encodingFormat: "float") so the returned vector length matches the model's real dimensionality; this avoids base64 interop quirks in proxies and mocks.

Optional provider integrations

Install the provider packages your agents use, as needed:

pnpm add @langchain/anthropic     # anthropic
pnpm add @langchain/google-genai  # google
pnpm add @langchain/mistralai     # mistral
pnpm add @langchain/groq          # groq
pnpm add @langchain/ollama        # ollama
pnpm add @langchain/xai           # xai
pnpm add @langchain/openrouter    # openrouter

Documentation

• Routes
• Memory
• Getting started

⭐ Star Dawn on GitHub · 📚 Read the docs · 💬 Ask in GitHub Discussions

License

MIT