Using Venice AI with OpenClaw

Prerequisites

• ✓An active Venice AI account with access to the API (sign up at https://venice.ai/).
• ✓A Venice API key in the `vapi_xxxxxxxxxxxx` format created under Settings > API Keys.
• ✓OpenClaw installed and available on your PATH so you can run the `openclaw` CLI.
• ✓Network access from your environment to `https://api.venice.ai/api/v1` over HTTPS.

Steps

1. 1

   Create a Venice AI API key

   Start by creating a Venice API key so OpenClaw can authenticate against the Venice API. Keep this key handy; you will either export it as an environment variable or pass it to the onboarding command.

   textCopy

   1.   Sign up at https://venice.ai/
   2.   Go to **Settings > API Keys > Create new key**
   3.   Copy your API key (format: `vapi_xxxxxxxxxxxx`)

2. 2

   Run interactive onboarding for Venice AI

   Use the OpenClaw onboarding flow to wire Venice into your setup with minimal manual config. This command prompts for your Venice API key (or uses `VENICE_API_KEY` if already set), discovers available models, and lets you choose a default Venice model.

   bashCopy

   openclaw onboard --auth-choice venice-api-key

3. 3

   Set the VENICE_API_KEY environment variable

   Export your Venice API key so OpenClaw and other tools can read it without hardcoding secrets in commands. This also lets the onboarding flow auto-detect your key and enables automatic model discovery from the Venice API.

   bashCopy

   export VENICE_API_KEY="vapi_xxxxxxxxxxxx"

4. 4

   Onboard Venice AI non-interactively

   If you script deployments or CI, use the non-interactive onboarding mode. This passes the Venice API key directly on the command line and configures the provider without prompts, which is useful for automated environments.

   bashCopy

   openclaw onboard --non-interactive \
     --auth-choice venice-api-key \
     --venice-api-key "vapi_xxxxxxxxxxxx"

5. 5

   Verify Venice AI is working with an agent call

   Once configured, run a quick agent call against a Venice model to confirm everything is wired correctly. This checks both authentication and model routing through OpenClaw.

   bashCopy

   openclaw agent --model venice/kimi-k2-5 --message "Hello, are you working?"

6. 6

   Choose and switch your default Venice model

   Pick a default Venice model that matches your use case, such as private reasoning or high-capability anonymized models. You can change this default at any time using the models CLI so your agents use the right Venice backend by default.

   bashCopy

   openclaw models set venice/kimi-k2-5
   openclaw models set venice/claude-opus-4-6

7. 7

   List available Venice models discovered by OpenClaw

   OpenClaw can list all Venice models it knows about, either from live discovery or the static catalog fallback. Use this to confirm that your key works and to see which private and anonymized models are currently available.

   bashCopy

   openclaw models list | grep venice

8. 8

   Configure Venice AI explicitly in your OpenClaw config

   For more control, define Venice as a provider in your OpenClaw config file. This lets you set the base URL, wire `VENICE_API_KEY`, and pin specific models and capabilities in a structured way.

   jsonCopy

   {
     env: { VENICE_API_KEY: "vapi_..." },
     agents: { defaults: { model: { primary: "venice/kimi-k2-5" } } },
     models: {
       mode: "merge",
       providers: {
         venice: {
           baseUrl: "https://api.venice.ai/api/v1",
           apiKey: "${VENICE_API_KEY}",
           api: "openai-completions",
           models: [
             {
               id: "kimi-k2-5",
               name: "Kimi K2.5",
               reasoning: true,
               input: ["text", "image"],
               cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
               contextWindow: 256000,
               maxTokens: 65536,
             },
           ],
         },
       },
     },
   }

9. 9

   Exercise different Venice models from the CLI

   Once Venice is wired up, call a variety of models to validate privacy modes, uncensored behavior, vision, and coding capabilities. These commands are also a quick way to sanity-check that streaming, vision, and other features behave as expected for each model.

   bashCopy

   # Use the default private model
   openclaw agent --model venice/kimi-k2-5 --message "Quick health check"
   
   # Use Claude Opus via Venice (anonymized)
   openclaw agent --model venice/claude-opus-4-6 --message "Summarize this task"
   
   # Use uncensored model
   openclaw agent --model venice/venice-uncensored --message "Draft options"
   
   # Use vision model with image
   openclaw agent --model venice/qwen3-vl-235b-a22b --message "Review attached image"
   
   # Use coding model
   openclaw agent --model venice/qwen3-coder-480b-a35b-instruct --message "Refactor this function"

Configuration

Troubleshooting

echo $VENICE_API_KEY
openclaw models list | grep venice

openclaw models list | grep venice

https://api.venice.ai/api/v1

openclaw models list

Frequently asked questions