Claude, GPT, Gemini, Grok, GLM and more in one workspace: research with sources, make images and video, draft docs and slides, and keep projects and memory together. Build on the same models with one API key.
Free to try, no account needed. Guest chat includes GPT-6.1 Sol, GPT-6 Luna and GLM 5.3 Derisked.
Illustrative preview
Compare these two launch plans and draft a one-page brief.
Plan B reaches customers two weeks sooner; Plan A carries less delivery risk. I've drafted the brief as a document, with a source for each claim.
Document · Launch briefOpenSwitched to GPT-6.1 Sol
Now make a cover image for it.
Models from
Two ways to use Vani
Use Vani Chat in your browser, or call the same models from your own code. Each has its own account and billing.
Vani Chat c.vani.ai
Chat, research and create with 37 models on one subscription, from a free plan up.
Vani API platform.vani.ai
One prepaid key for many models, through OpenAI- and Anthropic-compatible endpoints. No subscription needed.
Everything in one place
What you can do in Vani, from the first question to the finished deck, plus the same models on the API. Rolling out means being released now; work still in progress is under Coming soon.
Every model, one switcher
Change model at any point in a conversation, without starting over.
37 models · 11 labs
Long context
Long threads and large documents stay in view on models with very large context windows.
Up to 1M tokens · 33 models
Try it without an account
Guest chat works in your browser with no sign-up.
GPT-6.1 Sol, GPT-6 Luna and GLM 5.3 Derisked
Web search with sources
Vani searches the web first and cites the sources it used.
Bring your files
Attach PDFs, Word documents, CSV files and images, then ask about them.
Charts, diagrams and math
Replies can draw charts and diagrams, typeset formulas and offer quick choices you can click.
Learning mode
Vani guides you with questions and hints instead of handing over the answer.
Reasoning effort
Choose how hard a model thinks before it answers, on models that support it.
Low to Max
Fast mode
Switch on Fast for quicker replies from the models that offer it.
Fork from any message
Fork the conversation at any message to try another direction; the original stays intact.
Temporary chats
A conversation that is never saved, for quick questions.
Usage you can read
Weekly and billing-period progress with reset times, in plain numbers.
Actions on every message
Hover or tap a message to see when it was sent, and copy, edit, retry or fork it.
What's new, in the chatRolling out
A one-time note in the chat when something new ships, linked to the full post.
Images
Generate an image from a prompt, or upload one and describe the change.
GPT Image, Grok Imagine and Seedream
Video
Make a clip from text, animate an image, or set its first and last frames.
Seedance and Grok Imagine
Media library
Every image and video you make is kept in one place.
Vani Studio
Ready-made looks, camera moves and effects for your photos and videos.
Docs, Slides and Designs
Vani writes documents, slide decks and one-page designs you can edit, present, export and share with a link.
Projects with knowledge
Keep related chats, instructions and files together; Vani answers from the files and shows which ones it used.
Memory
Vani remembers useful details across chats. See, edit or turn it off at any time.
Deep research
A slower, deeper search that reads more sources and keeps each one beside the answer.
Scheduled tasks
Run a prompt on a schedule you choose; results arrive in your Updates.
Skills
Save instructions you reuse, import your GPTs and Gems, or add skills from the community gallery.
Connectors and MCP servers
Connect services and your own MCP servers so Vani can read from them and make changes you approve; it suggests one when it would help.
Monthly recap
A look back at your month: active days, the models you used most and what you made.
Your data, exportable
Export a copy of your account's data, or delete the account, from Settings.
Your coding tools, on Vani
Point Codex, Claude Code, Cursor, Cline or OpenCode at the Vani API, with a setup guide for each.
15 setup guides
Editors and SDKs
Guides for Xcode, VS Code, JetBrains, Microsoft Foundry and the official OpenAI and Anthropic SDKs.
Vani inside other assistants
Add Vani's MCP server to Claude, ChatGPT, Cursor, Claude Code, Codex or Gemini CLI to create images and video and run research there.
Code-aware project search
Add source files to a project; search matches names across styles, so computeBackoff finds compute_backoff.
Code & files
Vani runs Python on your files in a private, isolated sandbox to analyse data and make charts, spreadsheets, documents and slides.
OpenAI- and Anthropic-compatible
Keep your SDK and change the base URL: Chat Completions, Responses and Messages.
Prepaid, pay as you go
Add credit and pay per request. Eligible models cost 30% below their verified official rates.
Images and video on the API
Generate and edit images and make video clips with the same key.
Prompt caching
On models with a cached-input price, prompt tokens served from cache are billed at that lower rate.
MCP tools on the API
Give any model remote MCP tools through the Responses and Messages APIs.
A console with the detail
Usage analytics, the cost of every request, request logs, and your limits and tier.
Limits that grow with you
Rate-limit tiers follow your lifetime deposits; each tier raises your limits.
On a paid Vani Chat plan?
Create API keys billed to your plan in Developer settings (Plus and higher plans).
Vani Code Agent for VS Code
A coding agent in your editor: it edits files with a diff to review and runs commands only when you approve.
01Create
Make images with GPT Image, Grok Imagine and Seedream, and video with Seedance and Grok Imagine. Keep refining in the same thread; every result stays one click away.
01
Describe what you want and create it with leading image models.
02
Upload a photo or pick a generated image, then describe the change.
03
Turn an image into a video clip, or go straight from text to video.
04
Ask for changes to a video and get a new version in the same thread.
Every image and video is kept in your media library.
Included on every plan, Free too — generations use your plan's allowance.
How media worksVani StudioLive
Illustrations show the workflow; they are not model output.
02Work
Write documents and slides with Vani, keep each project's files and instructions together, and let routine work run on a schedule.
Vani writes documents, slide decks and one-page designs you can edit, present, export and share with a link.
Keep related chats, instructions and files together; Vani answers from the files and shows which ones it used.
Vani remembers useful details across chats. See, edit or turn it off at any time.
Run a prompt on a schedule you choose; results arrive in your Updates.
Save instructions you reuse, import your GPTs and Gems, or add skills from the community gallery.
Connect services and your own MCP servers so Vani can read from them and make changes you approve; it suggests one when it would help.
Illustrations show the product's wording; they are not screenshots.
03Code
Switch between frontier coding models in chat, point your coding agent at the Vani API, and bring Vani's tools into the assistants you already work in.
New · VS Code extension
A coding agent in VS Code, backed by the Vani API: chat, read and edit files with diff review, and run commands with your approval.
The extension is free. It uses your Vani API key: prepaid on the API platform, or billed to a Plus or higher plan.
Illustration.
Your coding tools, on Vani
Point Codex, Claude Code, Cursor, Cline or OpenCode at the Vani API, with a setup guide for each.
15 setup guides
Editors and SDKs
Guides for Xcode, VS Code, JetBrains, Microsoft Foundry and the official OpenAI and Anthropic SDKs.
Vani inside other assistants
Add Vani's MCP server to Claude, ChatGPT, Cursor, Claude Code, Codex or Gemini CLI to create images and video and run research there.
Code-aware project search
Add source files to a project; search matches names across styles, so computeBackoff finds compute_backoff.
Code & files
Vani runs Python on your files in a private, isolated sandbox to analyse data and make charts, spreadsheets, documents and slides.
BrowserComing soon
Vani opens and reads web pages for you, and asks before it submits a form.
DeployComing soon
Publish an app you build in chat as a live web app.
# Claude Code
$ claude mcp add --transport http vani https://mcp.vani.ai/mcp
# Codex
$ codex mcp add vani --url https://mcp.vani.ai/mcp
# Gemini CLI
$ gemini mcp add --transport http vani https://mcp.vani.ai/mcpCommands from the MCP setup guide; the sandbox card is an illustration.
04Models
Every plan, Free included, can use every model below. Higher plans add room to work, not a different list.
Free to try works in guest chat, without an accountDe-risked clearly labelled variants, explained below
05De-risked models
De-risked variants are designed for legitimate research, security, creative and technical work that consumer-oriented refusal layers often block.
Marked in red in chat, on pricing and in the API model list, so you always know when you are using one.
Pick one when your work calls for it. Standard models stay a click away in the same model switcher.
Use is governed by our Terms of Service, which prohibit illegal and harmful use. You are responsible for how you use what you generate.
zhipu/glm-5.3-deriskedFree to try in guest chat, and available on every plan.
06Plans and usage
See exactly how many tokens each plan includes for each model, per month and per week, before you subscribe. Cards show GLM-5.3.
Every plan includes every model in the catalogue · image and video generation · reasoning up to Max
Explore the workspace
$0/month
0.2× capacity
~35M tokens / month
~8.1M / week
with GLM-5.3For everyday work
$18/month
1× capacity
~166M tokens / month
~38M / week
with GLM-5.3For regular, heavier use
$78/month
4.8× capacity
~800M tokens / month
~184M / week
with GLM-5.3For long, focused sessions
$188/month
11.6× capacity
~1.9B tokens / month
~444M / week
with GLM-5.3For the heaviest workloads
$464/month
28.8× capacity
~4.8B tokens / month
~1.1B / week
with GLM-5.3Weekly limit in grey. The top models we offer by Artificial Analysis Intelligence Index, plus our top de-risked model; every model is on the pricing page.
| Model | Free$0/mo | Plus$18/mo | Pro$78/mo | Max$188/mo | Ultra$464/mo |
|---|---|---|---|---|---|
| Claude Opus 5.5 | ~13M~3M / week | ~62M~14M / week | ~300M~69M / week | ~725M~166M / week | ~1.8B~413M / week |
| Claude Sonnet 5.5 | ~26M~6.1M / week | ~125M~28M / week | ~600M~138M / week | ~1.4B~333M / week | ~3.6B~827M / week |
| Claude Fable 5.1 | ~6.5M~1.5M / week | ~31M~7.1M / week | ~150M~34M / week | ~362M~83M / week | ~900M~206M / week |
| GPT-6 Astra | ~5.2M~1.2M / week | ~25M~5.7M / week | ~120M~27M / week | ~290M~66M / week | ~720M~165M / week |
| GPT-6.1 Sol | ~17M~4M / week | ~83M~19M / week | ~400M~92M / week | ~966M~222M / week | ~2.4B~551M / week |
| Claude Opus 5 | ~13M~3M / week | ~62M~14M / week | ~300M~69M / week | ~725M~166M / week | ~1.8B~413M / week |
| GLM 5.3 DeriskedDe-risked | ~474K~110K / week | ~2.2M~518K / week | ~10M~2.4M / week | ~26M~6M / week | ~64M~14M / week |
Based on data from Artificial Analysis (artificialanalysis.ai), checked 13 Oct 2026. The ranking is Vani's and is not endorsed by Artificial Analysis.
Token figures are estimates from each plan's allowance and the model's rate, rounded down; they are not a guarantee. Weekly limits also apply, so one week is capped at the weekly figure. Actual usage depends on your input and output mix, context length and reasoning. Capacity is relative to Plus (1×). Estimated 14 Oct 2026, 14:16 UTC.
07Why Vani
One workspace for the models you use, limits you can read before you pay, and an API priced 30% below official rates on eligible models.
Claude, GPT, Gemini, Grok, GLM and more under one subscription, with one switcher — instead of a separate app for each.
See the modelsEvery plan publishes token estimates per model, monthly and weekly, and your dashboard shows progress and reset times.
Compare plansOn the API platform, eligible models are priced 30% below their verified official rates. Prepaid, with no subscription.
Explore the APIAn option for legitimate research, security, creative and technical work — always marked in red, so the choice is yours.
How it worksGuest chat with GPT-6.1 Sol, GPT-6 Luna and GLM 5.3 Derisked works without an account.
Start chattingA public status page, documentation, in-app support tickets, and Stripe billing — cancel your plan at any time.
View status08API platform
Prepaid credit, no subscription. Point your SDK or coding tool at https://api.vani.ai/v1 and choose any model in the catalogue.
Keep your SDK and change the base URL: Chat Completions, Responses and Messages.
/v1/chat/completions/v1/responses/v1/messages
Add credit and pay per request. Eligible models cost 30% below their verified official rates.
Generate and edit images and make video clips with the same key.
On models with a cached-input price, prompt tokens served from cache are billed at that lower rate.
Give any model remote MCP tools through the Responses and Messages APIs.
Usage analytics, the cost of every request, request logs, and your limits and tier.
Rate-limit tiers follow your lifetime deposits; each tier raises your limits.
Create keys billed to your plan in Developer settings (Plus and higher plans).
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.vani.ai/v1",
api_key=os.environ["VANI_API_KEY"], # a vani_ key
)
reply = client.chat.completions.create(
model="anthropic/claude-sonnet-5",
messages=[{"role": "user", "content": "Hello"}],
)
print(reply.choices[0].message.content)Set VANI_API_KEY to a vani_ key from the API platform.
Step-by-step setup for 15 popular clients and coding agents, with notes on what each one supports.
Also: Hermes Agent, Continue, Aider, Zed, Open WebUI, LibreChat, LobeHub / LobeChat, Cherry Studio and AnythingLLM. Browse all guides
09Coming soon
None of these is available yet, and we do not promise dates. Each one will appear in the app, and in the news, when it is ready.
Describe a multi-step job; Vani plans it and works through it while you are away, with a live timeline.
Coming soonConnected services that have their own interface can show it right in the conversation.
Coming soonVani opens and reads web pages for you, and asks before it submits a form.
Coming soonPublish an app you build in chat as a live web app.
Coming soonPublishers can submit an MCP server to the Connectors directory; each one is reviewed before it is listed.
Coming soon10FAQ
No. Open the chat and start typing — guest chat includes GPT-6.1 Sol, GPT-6 Luna and GLM 5.3 Derisked, on a free allowance shared by everyone trying Vani. Create a free account for your own allowance, saved history and the full model list.
Each plan includes an allowance measured in tokens, with monthly and weekly limits. For example, Plus includes about 166M GLM-5.3 tokens a month, and Ultra has 28.8× the capacity of Plus. The pricing page lists estimates for every model.
The catalogue currently has 37 models from 11 labs, including Claude, GPT, Gemini, Grok, GLM and more. Every plan can use every model; higher plans give you more capacity. You can switch models at any point in a conversation.
Yes, on every plan. Generate images, edit ones you upload or create, animate an image into video, and ask for changes to a video, all in the chat. Media uses your plan's allowance, and results are kept in your media library.
Yes. Choose Document, Slides or Design in the message box, or simply ask, and Vani writes it in a side panel. Edit it in place or ask Vani to change it, keep earlier versions, present slides full screen, export to PDF, Word, PowerPoint, Markdown or HTML, and share a read-only link.
It can. Memory keeps useful details from your conversations, and a reply that used one says so. You can see, edit or delete what Vani remembers, bring memories over from other assistants, or turn memory off in Settings → Personalization. Memory is never added to API requests.
A Vani Chat subscription covers the chat workspace, and on Plus and higher plans you can create API keys billed to your plan in Developer settings. The API platform at platform.vani.ai is separate and prepaid, with its own account: add credit, create vani_ keys and pay per request, with eligible models 30% below verified official rates. No subscription needed.
The API offers Chat Completions, Responses and Anthropic Messages endpoints, and the docs include setup guides for Codex, Claude Code, OpenCode, Cline, Cursor and other popular tools, each with notes on what that client supports.
Variants designed for legitimate research, security, creative and technical work that consumer-oriented refusal layers often block. They are marked in red everywhere, so choosing one is always deliberate, and their use is governed by our Terms of Service.
Plans are monthly subscriptions in US dollars, billed through Stripe. You can cancel at any time and keep access until the end of the period you have paid for. API credit on the API platform is prepaid and separate.
We are working on Background tasks, Interactive connectors, Browser, Deploy and Connector submissions. None of them is available yet and we do not give dates; the Coming soon section describes each one, and News announces it when it ships.
Not yet. Desktop apps for macOS, Windows and Linux are next, and a Vani CLI and mobile apps for iOS and Android are on the way. You can already use Vani in VS Code with our extension, Vani Code Agent (see Downloads), and in any modern browser.
The public status page at status.vani.ai shows the current state of each service. Signed-in users can open a support ticket from inside the app, and the docs cover setup and troubleshooting.
Vani · Begin
Start free in your browser — no account needed. Upgrade when you want more room, or build on the API.
We use cookies for analytics and to measure how well our campaigns are working. None of it is needed to run the site, and your conversations are never included. See our Privacy Policy.