https://github.com/cacheplane/dawnai
Build LangGraph agents like Next.js apps.
https://github.com/cacheplane/dawnai
agents ai developer-tools framework langchain langgraph typescript
Last synced: 6 days ago
JSON representation
Build LangGraph agents like Next.js apps.
- Host: GitHub
- URL: https://github.com/cacheplane/dawnai
- Owner: cacheplane
- License: mit
- Created: 2026-04-14T03:58:13.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2026-06-06T23:09:53.000Z (11 days ago)
- Last Synced: 2026-06-07T01:12:05.840Z (11 days ago)
- Topics: agents, ai, developer-tools, framework, langchain, langgraph, typescript
- Language: TypeScript
- Homepage: https://dawnai.org
- Size: 2.63 MB
- Stars: 84
- Watchers: 2
- Forks: 2
- Open Issues: 7
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
- Codeowners: .github/CODEOWNERS
- Security: SECURITY.md
- Support: SUPPORT.md
Awesome Lists containing this project
README
# Dawn
[](https://github.com/cacheplane/dawnai/actions/workflows/ci.yml)
[](https://github.com/cacheplane/dawnai/actions/workflows/scorecard.yml)
[](./LICENSE)
The TypeScript meta-framework for LangGraph. Author AI agents and workflows as filesystem routes, get end-to-end types and a local dev server for free, and deploy to LangSmith with one command.
## Why Dawn?
- **Kill the LangGraph boilerplate.** Export one `agent({ model, systemPrompt })` descriptor. Dawn discovers it, wires route-local tools into the generated graph, and emits a `langgraph.json` package ready for LangSmith.
- **Filesystem-routed agents.** Filesystem routes under `src/app/` — colocate state schemas, tools, middleware, and tests next to the route they belong to. No more ad-hoc folders.
- **A real local dev loop.** `dawn dev` runs your routes locally with LangSmith-style endpoints. Iterate in seconds, then verify the generated deployment artifact before shipping.
- **Typed end to end (TypeScript).** Route params, state, and tool I/O are generated as TypeScript types. `dawn verify` is your pre-deploy gate.
## Without Dawn / With Dawn
Same LangGraph deployment shape, less code to author.
### Without Dawn
```ts
// graph.ts
import { StateGraph, MessagesAnnotation, START, END } from "@langchain/langgraph"
import { ToolNode } from "@langchain/langgraph/prebuilt"
import { ChatOpenAI } from "@langchain/openai"
import { tool } from "@langchain/core/tools"
import { z } from "zod"
const greet = tool(async ({ name }) => `Hello, ${name}!`, {
name: "greet",
description: "Greet a user by name.",
schema: z.object({ name: z.string() }),
})
const model = new ChatOpenAI({ model: "gpt-4o-mini" }).bindTools([greet])
const tools = new ToolNode([greet])
async function callModel(state: typeof MessagesAnnotation.State) {
return { messages: [await model.invoke(state.messages)] }
}
function shouldContinue(state: typeof MessagesAnnotation.State) {
const last = state.messages.at(-1) as any
return last?.tool_calls?.length ? "tools" : END
}
export const graph = new StateGraph(MessagesAnnotation)
.addNode("agent", callModel)
.addNode("tools", tools)
.addEdge(START, "agent")
.addConditionalEdges("agent", shouldContinue, ["tools", END])
.addEdge("tools", "agent")
.compile()
```
```json
// langgraph.json
{
"dependencies": ["."],
"graphs": { "hello": "./graph.ts:graph" },
"node_version": "22",
"env": ".env"
}
```
### With Dawn
```ts
// src/app/(public)/hello/[tenant]/index.ts
import { agent } from "@dawn-ai/sdk"
export default agent({
model: "gpt-4o-mini",
systemPrompt: "You are a helpful assistant for the {tenant} organization.",
})
```
```ts
// src/app/(public)/hello/[tenant]/tools/greet.ts
export default async ({ name }: { name: string }) => `Hello, ${name}!`
```
`dawn build` emits the `langgraph.json` for you.
## Quickstart
1. Create a new app.
```bash
pnpm create dawn-ai-app my-dawn-app
cd my-dawn-app
pnpm install
```
2. Validate the app and generate types in one call.
```bash
pnpm exec dawn verify
```
3. Run the scaffolded route. The route path must be quoted because it contains `(`, `)`, and `[]`.
```bash
echo '{"tenant":"acme"}' | pnpm exec dawn run '/hello/[tenant]'
```
4. Optionally start the local runtime in one terminal and send the same route through `--url` from another terminal.
```bash
pnpm exec dawn dev --port 3001
echo '{"tenant":"acme"}' | pnpm exec dawn run '/hello/[tenant]' --url http://127.0.0.1:3001
```
## 30-Second Route
Dawn routes live under `src/app` and export one runtime entry. New agent routes should use the `agent()` descriptor from `@dawn-ai/sdk`; Dawn discovers the route, wires route-local tools into the generated graph, generates types, and produces a `langgraph.json` package for LangSmith.
```ts
import { agent } from "@dawn-ai/sdk"
export default agent({
model: "gpt-4o-mini",
systemPrompt: "You are a helpful assistant for the {tenant} organization.",
retry: { maxAttempts: 3, baseDelay: 250 },
})
```
Add `state.ts` for a route state schema, `tools/*.ts` for route-local tools, `middleware.ts` for access control, and `run.test.ts` for colocated scenarios.
The built-in `agent()` route materializes to a LangChain chat model. Dawn infers providers for known model families; set `provider` explicitly to one of the supported built-in provider ids for aliases, ambiguous model names, local models, or provider-router model ids. Raw `graph` and `chain` routes can still instantiate any provider directly.
---
⭐ [Star Dawn on GitHub](https://github.com/cacheplane/dawnai) · 📚 [Read the docs](https://dawnai.org/docs/getting-started) · 💬 [Ask in GitHub Discussions](https://github.com/cacheplane/dawnai/discussions)
## Learn more
- [Getting started](https://dawnai.org/docs/getting-started)
- [Routes](https://dawnai.org/docs/routes)
- [Tools](https://dawnai.org/docs/tools)
- [State](https://dawnai.org/docs/state)
- [CLI](https://dawnai.org/docs/cli)
- [Dev server](https://dawnai.org/docs/dev-server)
- [Testing](https://dawnai.org/docs/testing)
- [Deployment](https://dawnai.org/docs/deployment)
---
Contributions welcome — see [CONTRIBUTING.md](./CONTRIBUTING.md). Repo layout and dev commands in [CONTRIBUTORS.md](./CONTRIBUTORS.md). Security: [SECURITY.md](./SECURITY.md). Please follow the [Code of Conduct](./CODE_OF_CONDUCT.md).
## License
MIT. See [LICENSE](./LICENSE).