Venice
AI assistant that keeps chats on your device, with open-source and frontier models and plans from free to $200 a month.
Founded 2024
Some links on this page are affiliate links. If you buy through them we may earn a commission at no extra cost to you. This never affects our verdicts.
Venice is a privacy-first AI assistant founded by Erik Voorhees in 2024. It offers open-source models, some marketed as uncensored, and anonymized access to models from OpenAI, Anthropic and Google. It searches the web, reads uploaded documents and images, generates and edits images and holds voice conversations, on the web and on iOS and Android. Chat history stays on your device, and its terms say Venice will not train on your content and will require model providers not to.
The Free plan allows 10 text prompts and 15 image prompts a day on base models. Pro costs $18 a month, or $180 billed yearly, and adds unlimited text prompts, 1,000 images a day, document uploads and 100 credits ($1) a month. Pro Plus and Max ($68 and $200 a month) include 7,500 and 22,500 monthly credits for video, music, frontier models and the API. No team plan is listed.
Best for
People who want an AI assistant that keeps chats on their own device, under terms that rule out training on them, with open-source and frontier models and image and video generation in one account.
Pros
- Chat history stays on your device, and Venice says it does not store your prompts or model responses on its servers
- Its terms say Venice will not train on your content and will require its model providers not to
- Pro costs $18 a month, or $180 billed yearly, with unlimited text prompts and 1,000 images a day
- One account covers open-source models, anonymized GPT, Claude and Gemini models, and image, video and music generation
- Hardware-isolated (TEE) and end-to-end encrypted models on Pro and above
Cons
- Free plan is limited to 10 text prompts and 15 image prompts a day, without document or image uploads
- Chats do not sync between devices; moving them needs an encrypted backup (Pro) or a share link
- Anonymized GPT, Claude and Gemini models still send your prompts to those providers, which Venice says likely save them
- API use is pay-per-use on every plan, with no unmetered tier; Pro's allowance is 100 credits ($1) a month
- No team plan is listed, and subscriptions are generally non-refundable
Pricing
| Plan | Monthly | Yearly (12-month total) |
|---|---|---|
| Free10 text and 15 image prompts a day; base models; web search | Free Free plan | |
| Pro$180 billed yearly (save 16.7%); unlimited text prompts, 1,000 images a day, 100 credits a month | $18/mo | $180/yr ($15/mo) |
| Pro Plus$735 billed yearly (save 10%); 7,500 credits a month ($75), unused credits banked for 2 months | $68/mo | $735/yr ($61.25/mo) |
| Max$2,160 billed yearly (save 10%); 22,500 credits a month ($225), banked for 3 months | $200/mo | $2,160/yr ($180/mo) |
Prices checked on .
Features
In AI chat assistants
| Feature | Venice |
|---|---|
| Free tier | Yes · 10 text and 15 image prompts a day on base models |
| File upload | Yes · Paid plans: documents up to 10 MB (about 250,000 characters) and images for vision models; not on Free |
| Web search | Yes · All plans, incl. Free; Brave or anonymized Google; off in end-to-end encrypted mode |
| Image generation | Yes · 15 image prompts a day on Free, 1,000 on Pro; editing, upscaling and background removal on paid plans |
| Voice mode | Yes · Realtime voice conversations in the web and mobile apps; Pro 1 hour, Pro Plus 5, Max 20 per 30 days; no Free quota listed |
| Memory/projects | Yes · Memory stored on your device (Settings > Memory), off in end-to-end encrypted mode; chat folders, but no projects feature |
| API access | Add-on · OpenAI-compatible API, pay per use on every plan incl. Free; monthly allowance $1 on Pro, $75 Pro Plus, $225 Max |
| Mobile apps | Yes · iPhone, iPad and Android (Google Play or APK); App Store version limits model behavior |
| Team plan | No · No team or business plan listed on the pricing page |
| Training opt-out | Yes · No opt-out needed: terms say Venice won't train on user content and will require its model providers not to |