Build AI Chat Apps

For Documentation

The fastest way to create production-ready chat with your data apps.
Ingest, customize, then deploy with one line of code.

Backed BY
Y Combinator LogoCombinator
See how companies implement Mendable

Trusted by top companies

From small startups to Fortune 500 companies, we're proud to be a cutting edge platform for any 'chat with data' applications

From support bots to product copilots

Built for multiple use cases


Embed our smart search bar and decrease customer activation times.

Customer Support

Resolve 50% of support tickets instantly with our AI chatbot that lives in your help center.

Sales Enablement

Give your entire sales team instant technical knowledge, reducing wait-times, streamlining engagement and amplifying ROI.

Product Copilots

Speed up adoption times with personalized assistants that knows your product inside out.

Over 20+ data connectors

Start by connecting your data

Mendable offers managed ingestion through a simple online GUI and through our API. You can easily add, modify, or delete different types of sources.

Easily Teach Your Model

Customize your model

Customize base model properties

GPT-3.5-Turbo and GPT-4 are supported with a variety of base models coming soon

Training through answer correction

Correct the answers generated by the model and it will instantly learn from your feedback

Custom prompt edits

Edit the prompt to prevent hallucinations, maintain voice and format requirements

Keep your data always updated

Mendable reingestion process offers CRON jobs and webhooks to keep your data synced and always up to date

Support Link

Have customers redirected to your customer support link when the bot can't answer their questions

Privacy-first features

Mendable provides custom private, open source LLMs depending on your needs

React, Vanilla JS, API

Choose your component

Mendable provides a variety of components ranging from search bars, to chat bubbles, to full CLIs built on our API. Customize them or easily build your own

Mendable component
From zero to production in minutes

Deploy anywhere

Deploy Mendable internally, externally, or both. Our API allows you to send and query data from anywhere.

Saif Khan
Mendables integration on Nylas is a goldmine of data. Now, the product team has a direct source of user feedback, questions, and problems. It's amazing!
Explore your dashboard

Get insights from usage


Number of chat messages per month

Understand all interactions

Unravel your users' queries, track their interactions, customize responses, and monitor your product usage effortlessly.

  • Gain key insights into user queries
  • Monitor real-time product-user interactions
  • Fine-tune your model for optimized responses
  • Track and evaluate Mendable usage

Insights beyond conversations

Learn what your users are asking, how they are asking, and their satisfaction level with the answers. Teach the model based on the answers rating and improve the model's performance.

Our wall of love

Don't take our word for it

Empower your users with AI powered search

You are minutes away from supercharging your knowledge base

Frequently asked questions

If you have anything else you want to ask,reach out to us.

    • Is it free?

      We have a free plan that is limited to 100 chats a month. It is also free for certain Open source projects. Contact us to see if your project is eligible.

    • Do you train your AI model with my code?

      Currently, Mendable does not look at any of your repository's code. However, in the future we may add it. We will always give you the option to opt out of sharing your data.

    • How do I remove the Powered by Mendable?

      To remove the Powered by Mendable, you need to upgrade to an enterprise or custom plan. Contact us at and we can help you out.

    • How do I get an anon key?

      To get your anon key you need to sign up at and create a project. Then you can find your anon key in the API Keys section of the dashboard. Anon keys are used for client-side while API keys are used for server-side.

    • Which model does Mendable use?

      Mendable offers gpt-3.5-turbo and gpt-4. If you'd like a custom model, contact us and we can help you out.

    • Is GPT-4 pricing different?

      Yes, right now GPT-4 will cost 3 requests per message instead of 1 (gpt-3.5-turbo). That means that instead of 500 messages, you will get around 166 messages if you only use GPT-4.

    • Can you correct the AI response?

      Yes, Mendable offers a 'teach the model' functionality where you can correct the AI response and it will learn from it.

    • How can I integrate Mendable with my application?

      Check out the Mendable documentation here to better understand how you can start integrating.

    • Is it 100% accurate?

      Like Humans, AI will never be 100% accurate. So we can't assure you that every solution will be correct.

    • How do I cancel my subscription?

      Simply log into our platform, go to your account and click on "Open customer portal" button. There you will be able to cancel/modify it through Stripe.

    • How does Mendable work?

      Our application syncs with your documentation and support channels, then uses your docs and previously answered questions to suggest possible answers.

    • Are you open-source?

      We are starting to open source some of our technology, but currently most of it is not. However, we're strongly considering moving to a more open-source model. If you have input here, please message us