MiroMind API
Access MiroMind instantly with Puter.js, and add AI to any app in a few lines of code without backend or API keys.
// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
puter.ai.chat("Explain AI like I'm five!", {
model: "miromind/mirothinker-1-7-deepresearch"
}).then(response => {
console.log(response);
});
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
puter.ai.chat("Explain AI like I'm five!", {
model: "miromind/mirothinker-1-7-deepresearch"
}).then(response => {
console.log(response);
});
</script>
</body>
</html>
List of MiroMind Models
MiroThinker 1.7 DeepResearch
miromind/mirothinker-1-7-deepresearch
MiroThinker 1.7 DeepResearch is an open-source deep research agent from MiroMind AI, built on a 235B-parameter Qwen3-235B-A22B-Thinking-2507 mixture-of-experts architecture. It is designed for long-horizon web research, complex multi-step investigation, and prediction tasks rather than general chat. It scores 74.0 on BrowseComp, 75.3 on BrowseComp-ZH, 82.7 on GAIA-Val-165, and 42.9 on HLE-Text, and MiroMind reports it sets a new state of the art among open-source models on BrowseComp-ZH. The model supports a 256K-token context window and can issue up to 300 tool calls in a single task. MiroMind calls this approach interaction scaling, training the agent to handle deeper and more frequent tool interactions rather than relying only on larger parameter counts or longer context. Released under Apache 2.0, it's a fit for developers building autonomous research agents, web-browsing pipelines, or report-generation tools that need many sequential tool calls to reach a verified answer.
ChatMiroThinker 1.7 DeepResearch Mini
miromind/mirothinker-1-7-deepresearch-mini
MiroThinker 1.7 DeepResearch Mini is an open-source deep research agent from MiroMind AI, a 30B-parameter mixture-of-experts model built on the Qwen3-30B-A3B-Thinking-2507 architecture. It is the smaller sibling of MiroThinker 1.7 DeepResearch, aimed at the same long-horizon web research and multi-step investigation tasks at lower compute cost. It scores 72.3 on BrowseComp-ZH, which MiroMind reports as a new state of the art among open-source models at this parameter scale. Like the full-size model, it supports a 256K-token context window and can issue up to 300 tool calls per task, using the same interaction-scaling training approach that emphasizes deeper and more frequent tool use over simply adding parameters. Released under Apache 2.0, it suits developers who want MiroThinker's research-agent behavior in a lower-cost, faster model for high-volume or latency-sensitive deep-research API calls.
Frequently Asked Questions
The MiroMind API gives you access to models for AI chat. Through Puter.js, you can start using MiroMind models instantly with zero setup or configuration.
Puter.js supports a variety of MiroMind models, including MiroThinker 1.7 DeepResearch and MiroThinker 1.7 DeepResearch Mini. Find all AI models supported by Puter.js in the AI model list.
With the User-Pays model, users cover their own AI costs through their Puter account. This means you can build apps without worrying about infrastructure expenses.
Puter.js is a JavaScript library that provides access to AI, storage, and other cloud services directly from a single API. It handles authentication, infrastructure, and scaling so you can focus on building your app.
Yes — the MiroMind API through Puter.js works with any JavaScript framework, Node.js, or plain HTML. Just include the library and start building. See the documentation for more details.