TL;DR: We (Sage, a nonprofit) built Talleyrand, a free, open-source AI workspace for thinking through hard questions. Your questions are kept in a tree, not a linear chat. After each answer it suggests the next questions and learns from your reactions. Try it at talleyrand.app.
Why we built it
I like to discuss complex topics with LLMs, but their linear chat is a bad place to think through hard questions. Every good answer raises three new questions. I ask one, then scroll back and forth to remember the other two. Twenty messages in, I've forgotten some of the questions I wanted to dig into. I start a new chat and it forgets everything, so I re-read what I already know. And I can't just click "yeah, I know this, skip it" or "off topic".
So we built Talleyrand to keep the questions in a tree.
Four things it does that a chat doesn't:
1. Start from a mess, not a blank page. Don't know where to start with a complex topic? Type or dictate what you know, what you want to know, who you are, your doubts, half-baked thoughts. Talleyrand summarizes it into a short brief and generates your first questions.
2. Branch out from any sentence. See something you want to know more about? Select it and ask. You can also mark the selection as an insight, so the LLM will know it’s interesting to you.
3. One tree for the whole case. Every LLM prompt is built from the brief, the other questions, your highlights, your reactions and what you’ve read, so the answers are personalized to you. The tree shows what's waiting, what you've read, and what's settled. The LLM sees that too.
4. It suggests questions and learns from you. After each answer it suggests follow-up and "big-picture" questions that challenge your framing. Accept them, or decline and say why. Every heart, highlight and decline improves the next suggestion.
I personally use it for questions which I expect to take more than three questions in a dialog with an LLM, which is mostly AI research these days.
Sign in with Google, bring your own OpenAI key. If you want to use Claude, add an Anthropic key.
Keys are stored in your browser only and sent with each request. They're never stored on the server. You pay your provider for tokens.
Self-hostable with Docker. The README has the instruction. The only setup work is Google OAuth credentials, which I know is annoying, but you can ask Claude to remove that if you need to.
Feedback
Comments here, GitHub issues, email talleyrand@sage-future.org, or the Sage Discord. I read everything, and I'm happy to iterate on your feedback and make it easier to use.
TL;DR: We (Sage, a nonprofit) built Talleyrand, a free, open-source AI workspace for thinking through hard questions. Your questions are kept in a tree, not a linear chat. After each answer it suggests the next questions and learns from your reactions. Try it at talleyrand.app.
Why we built it
I like to discuss complex topics with LLMs, but their linear chat is a bad place to think through hard questions. Every good answer raises three new questions. I ask one, then scroll back and forth to remember the other two. Twenty messages in, I've forgotten some of the questions I wanted to dig into. I start a new chat and it forgets everything, so I re-read what I already know. And I can't just click "yeah, I know this, skip it" or "off topic".
So we built Talleyrand to keep the questions in a tree.
Four things it does that a chat doesn't:
1. Start from a mess, not a blank page. Don't know where to start with a complex topic? Type or dictate what you know, what you want to know, who you are, your doubts, half-baked thoughts. Talleyrand summarizes it into a short brief and generates your first questions.
2. Branch out from any sentence. See something you want to know more about? Select it and ask. You can also mark the selection as an insight, so the LLM will know it’s interesting to you.
3. One tree for the whole case. Every LLM prompt is built from the brief, the other questions, your highlights, your reactions and what you’ve read, so the answers are personalized to you. The tree shows what's waiting, what you've read, and what's settled. The LLM sees that too.
4. It suggests questions and learns from you. After each answer it suggests follow-up and "big-picture" questions that challenge your framing. Accept them, or decline and say why. Every heart, highlight and decline improves the next suggestion.
I personally use it for questions which I expect to take more than three questions in a dialog with an LLM, which is mostly AI research these days.
Cost, keys and data
Feedback
Comments here, GitHub issues, email talleyrand@sage-future.org, or the Sage Discord. I read everything, and I'm happy to iterate on your feedback and make it easier to use.
Talleyrand is built at Sage, the nonprofit behind AI Digest, AI Village, Fatebook and Quantified Intuitions.