Venice API

Developer guides

From first request to production agents.

Start with an OpenAI-compatible API call, then add Venice-specific privacy, tools, media generation, and wallet-funded agent workflows.

Get an API key →Migrate from OpenAI

Start fast

Make your first useful request

The shortest path from account setup to a working integration.

01Create an API keyGenerate a key in the Venice dashboard and store it in your environment.02Point your client at VeniceKeep your OpenAI-compatible code and change the base URL.03Test the API shapeUse Postman to inspect auth, request bodies, responses, and errors.

Build paths

Choose what you are building

Follow the path that matches your app, then jump deeper when you need a specific capability.

Text & chatBuild reliable chat workflowsStructured responses, reasoning models, file inputs, prompt caching, and TEE/E2EE privacy controls.Start with structured responses →MediaGenerate images and videoImage generation, editing, video jobs, reference-to-video, upscaling, and consent-aware media flows.Explore media guides →AgentsConnect Venice to agentsMessaging bots, coding agents, MCP tools, runtime skills, and crypto-aware agent workflows.Open the agents hub →PaymentsFund requests your wayUse API credits, DIEM staking allowances, x402 wallet payments, and Crypto RPC endpoints.Read the x402 guide →

Developer journey

Move from prototype to shipped product

01PrototypePick a model and shape the first conversation loop.02Add contextAttach files, search the web, and structure outputs.03IntegrateExpose Venice through MCP servers, skills, SDKs, and app frameworks.04OptimizeReduce latency and cost with caching, privacy modes, and operational patterns.

Keep learning

Reference patterns and project walkthroughs

Coding toolsUse Venice models in Cursor, Claude Code, OpenCode, and Codex CLI.SDKs & frameworksBuild with LangChain, Vercel AI SDK, and CrewAI.Project walkthroughsBuild private RAG, research, and code review agents end-to-end.

Agent quickstartAfter you create a key

Build a private research agent

Use any OpenAI-compatible client, point it at Venice, then turn on Venice-native tools like server-side web search with venice_parameters.

1export VENICE_API_KEY=“your-api-key”

2npm install openai

Explore agent workflows →

Generating an API KeyCreate a Venice API key from the dashboard, store the bearer token securely, and verify it works with a first authenticated chat completions request.

⌘I