Skip to main content
MITHUNAI is an AI knowledge and answer platform. Its core, MITHUNAI RAG, reads your documentation and other knowledge sources and answers questions from them with citations to the passages it used, or says it does not know when your content does not cover the question. You put those answers on your website, in your own product over the HTTP API, and in AI tools over MCP. Beside it sit AI search visibility, AI employees, Chat, Agents and the Gateway, and one account signs you in to all of them. Search every guide and the API reference with Ctrl K (⌘ K on a Mac), or start from what you do.

​
Start with your role

I build with the API

From an empty organisation to a cited answer with the HTTP API alone.

I run our organisation

Every step from a first look to a live assistant your team uses.

I lead support

Put Ask AI on your website or help centre with one line.

​
The products

MITHUNAI RAG

Connect your knowledge, answer with citations, and deploy through the widget, the HTTP API and MCP.

AI search visibility

Beta See how AI engines answer your buyers’ questions: which companies they name and which pages they cite.

AI employees

Beta Support, IT, DevOps, marketing, sales and finance employees that ask a person before they act.

MITHUNAI Chat

A chat workspace for your team, with your knowledge and your assistants in every thread.

MITHUNAI Agents

Beta Agents that plan over several steps and call tools. Access on request.

MITHUNAI Gateway

One OpenAI-compatible endpoint in front of your model providers, with routing and failover.

​
Who each product is for, and its status

Product status lists every capability and its status, AI search visibility included: it is in beta, switched off by default.

​
How MITHUNAI RAG works

1

Connect your knowledge

Add the sources your assistant should answer from in Sources. MITHUNAI fetches each one, extracts its text and indexes it in the background. Confluence, Notion, Jira, Zendesk tickets and private GitHub repositories read with a token your organisation saves once as a connection. See Knowledge sources.
2

Configure an assistant

Choose the knowledge an assistant answers from, its model and the guidance that sets its tone in Customizations, then question it in the Playground. See Create and configure an assistant.
3

Deploy it

Put the assistant on your website with the widget, call it from your backend with an API key, or connect Claude and your IDE over MCP. See Deploy.
4

Analyze and improve

Read what people asked, what they ask most, which questions your content could not answer and which sources the answers cite, and see how much of your monthly answer allowance is used. See Analyze.
Every answer follows the same rule on every channel: it is written from passages retrieved from your content, it cites them, and when the evidence is not strong enough the assistant abstains instead of guessing. How answers work explains each step.

​
Build with the API

API reference

Every endpoint, with its request and response fields.

Authentication

Send your API key as a bearer token, and see which routes need a console session.

​
What’s new

  • 11 November 2026: a Microsoft Teams bot, Expensify for the Finance employee, and four Beta features that stay off until your deployment switches them on.
  • 7 November 2026: a Slack bot, proposed answers for coverage gaps, approval rules, and AI employees that work in more of your tools.
  • 2 November 2026: features that are still being worked on are labelled Beta.
Every release, with what changed and what is still in beta, is in the changelog.

​
One account for every product

Your MITHUNAI account and your organisation belong to the platform, not to any one product. You sign in once on the console, and every other product lets you straight in with your organisation and role; signing out of one signs you out of all of them. Chat works this way today, and the rest is in beta: see Sign in once. How the products fit together explains what each one shares with the others.

​
Getting access

There is no self-service sign-up: you request access, or accept an invitation, and a person sets up your organisation. To evaluate MITHUNAI first, try the public playground or book a demo. If your organisation already uses it, ask its Owner or an Administrator to invite you by email. Your path to a live assistant shows every step after that. Forgotten your password? Use Forgot password? on the sign-in screen; see Password reset.

​
Trust and support

Security

How each organisation’s data is isolated, and what happens to your content.

Product status

What is available, what is in beta and what is planned.

Troubleshooting

Symptoms and fixes for answers, ingestion, signing in, the API and the widget.

Changelog

What changed in each release.
Last modified on November 12, 2026