Free multi-model AI chat in the browser with no sign-up, an open-source code base and an OpenAI-style API; Premium adds more models for $3.
Use it to try several newer models for free without registering, or to get a cheap API key to test with. Avoid it for private work: prompts go through small third-party routers, and no company stands behind it. Its GitHub repository was created in August 2026 and shows over 1,500 stars.
A no-signup chat that lets a visitor try several newer models free and then turn the same page into an OpenAI-style API.
The code is public and the models belong to other companies, so a user can leave at no cost and only saved Projects stay behind.
Main trade-off
Prompts pass through small third-party routers, with no company, status page or support contract behind the service.
| Plan | Monthly price | What it includes |
|---|---|---|
| Free | Free | 1,000,000 tokens per hour, Up to 30 messages per minute, All free models |
| Premium | $3/mo | Five times the Free plan's hourly tokens, Same message rate as Free, Premium models, also over the API |
Premium access is granted by hand rather than through an online checkout.
Ptero is a free multi-model AI chatbot that runs in the browser without an account. It lists nine third-party models, keeps chats outside Projects in browser storage, adds a code editor and an OpenAI-style API, and publishes its code on GitHub as a WordPress plugin. A paid Premium plan adds two models and more hourly tokens.
No sign-up chat
Pick a display name and start; the guest identity and chats outside Projects live in your browser's local storage.
Nine models in one picker
Seven free and two Premium models from third-party providers, switchable inside a chat, with a thumbs up or down tallied per model.
Projects with a code editor
Sign in with GitHub to keep chats and cloud files together, edit files in a Monaco editor and import a public GitHub repository.
OpenAI-style API
Personal API keys, created after GitHub verification, work against a chat and a streaming endpoint with the same hourly token quotas as the web app.
Open-source code
The code base is public on GitHub as a WordPress plugin, and the site links to it before you start chatting.
Interface in 22 languages
The language switcher covers Persian, Arabic, Chinese, Spanish, Hindi, Japanese, Korean and others.
| Tool | Price from | Free plan | Runs on | Rating |
|---|---|---|---|---|
| $3/mo | Yes | Web | 3.4 | |
| $4.99/mo | Yes | Web · iOS · Android · Mac · Windows | 4.0 | |
| $20/mo | Yes | Web · iOS · Mac · Windows | 3.7 | |
| $3/mo | Yes | Web · iOS · Android · Mac · Windows | 4.5 |
Open source: the repository LICENSE file is Apache-2.0, while its README says GPL v2 or later
The Free plan costs $0 and covers the free models. Premium is shown at $3, with 5,000,000 tokens an hour and the Premium models, and the page does not say how often it is charged. You get Premium by asking on the Discord server or by email, and the operator sets the account up by hand.
No. You pick a display name and the site stores a random guest token in your browser, and your hourly usage quota is tied to that identity and device. A GitHub sign-in is needed only for Projects and for creating API keys.
Chats outside Projects stay in your browser's local storage, and the privacy policy says they are not stored in the database. Chats inside a Project, its title and its files are saved on the operator's servers and deleted after 90 days without opening Projects. Clearing browser data removes the local chats.
The privacy policy lists hcnsec, kiraai, minirouter, river, runware and voyage as possible recipients, depending on the model you pick. Each handles the data under its own practices, and if a model fails the site may retry with another model, which can send the conversation to a different provider.
Yes, with a personal API key created after GitHub verification, sent as a Bearer token to the chat endpoint at ptero.pro/wp-json/mlp/v1/chat. Premium models need the Premium plan. The in-app guide shows placeholder addresses, and an unauthenticated test request on 22 November 2026 returned HTTP 403 asking for a valid WordPress nonce.
Ptero is a free AI chat that puts nine models from several providers behind one web page, with no account needed: a guest name and a random token are kept in the browser, and chats outside Projects stay there. The page lists seven free models and two Premium ones at the time of reading, among them Qwen, GLM, MiniMax and DeepSeek V4 Flash, plus a model called GPT 6.1 Terra. The operator gives no company name and publishes the code on GitHub as a WordPress plugin, with a licence file that reads Apache 2.0 while the README says GPL v2 or later, and the site offers 22 interface languages.
Beyond chat it has Projects, which need a GitHub sign-in: a project keeps chats and up to 35 files of 15 MB each on the server, has a code editor, and can import a public GitHub repository. A developer panel issues API keys for an OpenAI-style endpoint once GitHub verification is done, and there is an MCP entry in the menu. Usage is metered in tokens per hour for each guest identity and device.
It suits someone who wants to try several of the newer open and Chinese-lab models without paying or registering, a student who wants a free chat with a code editor, or a developer who wants a cheap API key to test against.
It is a poor fit for confidential work. Prompts go through the operator's server to small third-party routers rather than to the model makers, and the privacy policy leaves how those routers handle them to their own practices. It is also a poor fit for anyone who needs a company behind the service, a status page or a support contract. For a bigger catalogue with named labs look at Poe or Bearly, and for a pure API gateway at OpenRouter. A single maker's own free chat such as DeepSeek or Qwen removes the middle layer. More are listed under AI chatbots.
Product facts on this page (pricing, plans, integrations, compliance) are read from Ptero's own pages and documentation. The verdict is Hokai's reading of them, not the vendor's claim.