Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.typingmind.com/llms.txt

Use this file to discover all available pages before exploring further.

TypingMind is a unified chat interface for working with large language models — connect your own API keys and access ChatGPT, Claude, Gemini, and more in one organized workspace.

​
What You Can Do

Chat & Organize

Folders, tags, forking, multi-tab, and full-text search across all conversations.

Any AI Model

OpenAI, Anthropic, Google, Mistral, Grok, Azure, and more — all in one place.

Prompts & Agents

Build a reusable prompt library and custom AI assistants with training data.

Knowledge Base

Upload files or connect Google Drive, Notion, GitHub, and more for RAG-powered answers.

Plugins & MCP

Web search, image generation, deep research, custom tools, and MCP server integrations.

Privacy First

All chats stored locally by default. Optional cloud sync across devices.

​
Setup Guide

​
Step 1 — Activate your license & API keys

​
Activate your License Key

You can use TypingMind without a license key, but advanced features require one:
  • Multi-model parallel chats
  • Plugins (Web Search, Image Generator, etc.)
  • Projects and Artifacts
  • And more
Compare license plans to find the right tier. Once purchased, your license is automatically activated on your device.
License key activation screen To manage or retrieve your license key later, see Manage License & Devices. License management screen

​
Add your API Keys

TypingMind supports models from all major providers. You’ll need to obtain an API key directly from each provider you want to use:
ProviderModels
OpenAIGPT-4o, o1, GPT-4 Mini, GPT-3.5, and more
AnthropicClaude 3.5 Sonnet, Claude 3 family, and more
GoogleGemini 1.5 Flash, Gemini 1.5 Pro, and more
OthersMistral, DeepSeek, Grok, and more
One API key per provider unlocks all models from that provider. You do not need a ChatGPT Plus or Claude AI subscription — the API is a separate product.
To add your keys: go to Models in the left sidebar → click the gear icon next to a provider → enter your key → enable the models you want. API key setup screen Enabling models For a detailed walkthrough, see Manage & Connect AI Models.

​
Step 2 — Back up your data (optional)

Your chats are saved locally in your browser by default — nothing is sent to TypingMind’s servers unless you opt in. To protect against data loss and sync across devices, sign in with your email to enable TypingMind Cloud: Cloud sync login screen Cloud sync includes:
  • Chats, prompts, agents, plugins, API keys, and settings
  • Up to 50 MB free storage (expandable)
  • Seamless sync across all your devices
Cloud sync overview You can disable sync at any time by logging out, or connect your own cloud storage via TypingMind Extensions. Learn more in Cloud Sync & Backup.

​
Step 3 — Start your first conversation

Once your API keys are set up, select a model and start chatting. You can only use models for which you’ve added an API key.
Start with clear, specific instructions. If a response misses the mark, refine your prompt — more detail almost always yields better results.
First conversation screen

​
Step 4 — Set global AI behavior with System Instructions

System Instructions define how the AI behaves across your entire session — its tone, personality, and scope. Set them in Models → Global Settings.
Context length is the maximum number of tokens (words/characters) an AI model can hold in memory during a conversation. Once the limit is reached, older messages are dropped.System Instructions are persistent — the model always follows them, even after earlier messages fall out of context. For reference, GPT-4o supports up to 128,000 tokens.
System instructions screen Learn more: System Instruction

​
Step 5 — Get higher-quality responses

Save and reuse prompts for recurring workflows. Access your library from the Prompts menu in the left sidebar — use prebuilt prompts or create your own.Prompt libraryTip: Be specific. “Explain machine learning basics with examples of supervised and unsupervised learning” will outperform “Explain machine learning” every time.Learn more: Use Prompt Library
Attach files directly to a conversation using the attach icon or drag-and-drop. The AI reads the content and can summarize, explain, or answer questions based on it.File upload in chatTip: Well-structured, relevant documents yield the most accurate results — ideal for analyzing reports or summarizing long texts.Learn more: Chat with Files or Videos
Plugins give the AI access to the internet, image generation, code rendering, and more. Find them in the Plugins section of the left panel.Popular plugins include:You can also build Custom Plugins to connect your own APIs or MCP servers.Plugins panelCustom plugins
The Knowledge Base feature lets you upload files or connect external sources (Google Drive, GitHub, Notion, websites) so the AI can retrieve relevant context when answering your questions.Knowledge base screenTo get started: click the KB tab → Add Data Source → upload files or connect a source.Learn more at RAG Knowledge Base
AI Agents are purpose-built assistants with their own model, instructions, plugins, and training data. Create them from the Agents section in the left panel.Each agent can have:
  • A dedicated base model (e.g. GPT-4o)
  • Plugins for web search, image generation, or custom tools
  • A connected knowledge base or uploaded training files
  • Few-shot examples to guide behavior AI agents overview
Great for recurring, specialized tasks — like a customer support agent trained on your company FAQs.AI agent exampleLearn more: AI Agents Overview

​
Step 6 — Organize your workspace

Your full chat history lives in the left panel. Keep it tidy by:
  • Dragging chats into folders
  • Pinning important conversations to the top
  • Deleting stale chats
Chat organization For longer-term work, use Projects — containers that group related chats with shared model settings, instructions, and agents. Projects view Learn more: Organize Chats

​
Step 7 — Share your conversations

Click the Share button above the message area to export or share any chat in your preferred format:
  • 🔗 Secret link
  • 📝 Markdown
  • 📦 JSON
  • 🌐 HTML Webpage
  • 📄 PDF
Share chat options