Skip to main content

Webhook Pipeline

View as Markdown

Webhook Pipeline

The simplest useful pipeline: an HTTP endpoint that accepts a question, sends it to an LLM, and returns the answer. This is the "hello world" of RocketRide — a good starting point before adding vector stores, preprocessing, or agents.

The pipeline​

Save this as chat.pipe:

{
  "nodes": [
    {
      "id": "source_1",
      "provider": "webhook"
    },
    {
      "id": "llm_1",
      "provider": "llm_openai",
      "config": {
        "profile": "openai-4o-mini",
        "apikey": "${OPENAI_API_KEY}"
      },
      "input": [
        { "lane": "questions", "from": "source_1" }
      ]
    },
    {
      "id": "target_1",
      "provider": "response",
      "input": [
        { "lane": "answers", "from": "llm_1" }
      ]
    }
  ]
}

Three nodes, one connection each. The webhook source receives the request, the llm_openai node answers it, and the response target returns the result.

Start the pipeline​

export OPENAI_API_KEY=sk-...

rocketride start --pipeline ./chat.pipe

Output:

Webhook ready - system is ready to accept requests
  URL:  http://localhost:5567/task/data
  Auth: <public-auth-key>
  Token: <private-token>

Keep the terminal open — the pipeline runs until you stop it.

Send a question​

curl -X POST http://localhost:5567/task/data \
  -H "Authorization: Bearer <public-auth-key>" \
  -H "Content-Type: text/plain" \
  -d "What is the capital of France?"

Response:

{ "answer": "The capital of France is Paris." }

Use the Chat UI variant​

Swap webhook for chat to get a browser-based chat window instead of a raw HTTP endpoint, no curl required:

{ "id": "source_1", "provider": "chat" }

Restart the pipeline. The engine prints a chat URL:

Chat ready - system is ready to accept questions
  URL: http://localhost:5567/chat?auth=<public-auth-key>

Open that URL in a browser and start typing. The ?auth= query parameter is a convenience for the browser; the endpoint also accepts the key in an Authorization: Bearer <public-auth-key> header, the same scheme the webhook endpoint uses.

Add a system prompt​

Give the LLM a persona by adding a prompt node:

{
  "id": "prompt_1",
  "provider": "prompt",
  "config": {
    "text": "You are a helpful assistant that answers questions concisely."
  },
  "input": [
    { "lane": "questions", "from": "source_1" }
  ]
}

Then wire llm_1 to receive from prompt_1 instead of source_1.

Next steps​