Flows Examples

Example flows for Flows by Lit Protocol — the publish-and-pay platform for code that runs in a Trusted Execution Environment (TEE).

Publish a JavaScript function, set a price, and agents pay per request via x402, credits, or card. No servers to manage, no infra to maintain.

Examples

Flow	Description	Secrets	Price
oracle.js	Fetch a URL and sign the response with the flow's vault PKP	None	5c
model-council.js	Query 4 LLMs in parallel via OpenRouter, synthesize consensus	OPENROUTER_API_KEY	90c
embeddings.js	Generate OpenAI text embeddings from text or PDF	OPENAI_API_KEY	2c
pdf-extract.js	Extract text from a PDF using opendataloader-pdf	None	2c
lob-postcard.js	Send a physical postcard via the Lob mail API	LOB_SECRET_API_KEY	$1.00
non-us-proxy.js	Get a non-US (Mexico) proxy from a managed pool	WEBSHARE_API_KEY	2c
gmail-read.js	List recent emails (requires Gmail OAuth connection)	None (OAuth)	2c
gmail-reply.js	Send email replies (requires Gmail OAuth connection)	None (OAuth)	2c

Getting Started

Install the CLI

npm i -g @lit-protocol/flows
flows login

Publish a flow

# Simple flow — no secrets
flows publish oracle.js --name "My Oracle" --price 5

# Flow with a secret
flows publish embeddings.js --name "Embeddings" --price 2
flows secrets set embeddings OPENAI_API_KEY "sk-..."

# Flow with OAuth
flows publish gmail-read.js --name "Gmail Reader" --price 2 --connections gmail

Invoke a flow

# As a consumer
flows invoke embeddings --params '{"input": ["What is Lit Protocol?"]}'

# PDF mode
flows invoke embeddings --params '{"pdf_base64": "'$(base64 -i doc.pdf)'"}'

Use via API

curl -X POST https://flows.litprotocol.com/api/flows/embeddings/invoke \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"params": {"input": ["What is Lit Protocol?"]}}'

Every flow is also an MCP server at https://flows.litprotocol.com/mcp/<slug> — add it to Claude, Cursor, or any MCP-compatible agent.

How Flows Work

1. You write JavaScript. Your code runs inside async function main(params) { ... } in Lit Protocol's TEE.
2. Secrets stay secret. API keys are encrypted and only decrypted inside the TEE. Consumers never see them — they just pay and get results.
3. Every flow gets a vault. A programmable key pair (PKP) tied to your flow, useful for signing data or on-chain actions.
4. Consumers pay per request. Via x402 (USDC on Base), platform credits (Stripe), or card (Machine Payments Protocol). No accounts or API keys needed for x402.
5. Code is verifiable. The source is hashed to an IPFS CID at publish time. The TEE verifies the CID before execution — nobody can swap your code.

PDF Extraction Service

The embeddings.js and pdf-extract.js flows use a lightweight microservice at services/pdf-extract/ for PDF text extraction. It wraps opendataloader-pdf (a Java-based PDF extraction engine) in a thin Express server deployed to Fly.io.

The service scales to zero when idle — no cost when not in use.

Documentation

• Full docs
• CLI reference
• API reference
• Publishing guide
• Secrets & encryption

License

MIT

---

This repo is auto-synced from the main Flows repository. To contribute, open an issue or PR here.