Werner F.
I like TypingMind
I like the fact that I have different AIs.

TypingMind is an all-in-one AI workspace for using leading AI models from OpenAI, Anthropic Claude, Google Gemini, xAI Grok, DeepSeek, Perplexity, OpenRouter, Azure OpenAI, and more — all from one interface.Supports AI Agents, MCP servers, plugins, knowledge bases, document and image uploads, web search, artifacts, voice features, team collaboration, and self-hosted enterprise deployments.
Typing Mind (typingmind.com) has 672 customer reviews on Five Star Index, with an average rating of 4.6 out of 5.
Is this your business?
Correct your details and reply to customers in public. Both are free.
Claim this listingShowing 120 reviews with 4 stars · Clear filters
Werner F.
I like the fact that I have different AIs.
SergeD
Nice UI. Great access to several AI models via OpenRouter and excellent support for custom APIs/MCP tools. Unfortunately, it lacks native support for advanced agent frameworks—such as structured identity/personality profiles (like SOUL.md)—and complex context or long-term memory management.
Samy
I had 1 big freeze very frustrating when TypingMind tried to summarize my long context... Also I don't like the fact that changing the picture of an agent don't change the picture of a chat already started with this agent. I also would like a way to switch between sending all the context to sending only the X last message with less click. But that's not big problem, it mostly works fine.
NA
The keyboard shortcut to quickly delete chats is missing. Missing killer feature for me - chat branches with visualization
BobCo R.
it's good but a little API cost too much for me.
RossR
All mostly good. Things are still too fussy to set up sometimes and I hope we get more options for chat appearance such as fonts, etc. But the concept is great and when it works well it’s pretty amazing.
NoUser
easy to operate.
Eric A.
I love the plugins and comparing multiple llms. GPT hasn't been working on multiple llm mode. Also, it would be nice to be able to have parallel llm's be able to see each others results with a button. Also I would love a knowledge base with the ability to see more than 5 vector chunks at a time, as some of my documents had 300 vector chunks. Thanks again for all your work, I'm really enjoying it!