Get help with Symbol

Symbol is an intentional knowledge base for AI context. You save knowledge as structured capsules, organised into types with AI guidance. Any MCP-compatible AI tool loads those capsules by reference, giving the AI exactly what you chose to share.

If you work with AI regularly, you've probably noticed that valuable context gets lost inside chat histories. Symbol solves this by letting you save that context as structured capsules, organised into types with AI guidance. Whether you're a developer, designer, researcher, or support agent, Symbol gives your AI the right context every time.

You save knowledge as capsules, organised into types (e.g., Solutions, Decisions, Prompts). Each type carries guidance that tells AI how to structure and interpret capsules within it. Capsules get human-friendly reference IDs like @docker/3, so you can reference one capsule from another without copy-pasting. The MCP interface is the primary way to interact with Symbol: saving and loading context directly from AI conversations.

Sign up for an account, then connect Symbol to your AI assistant via the MCP integration. Start by creating a few Types that match your workflow (e.g., "Solutions", "Meeting Notes", "Code Snippets"), then save your first Capsule from an AI conversation. You can also create Capsules directly from the web interface.

AI creates a capsule that you are mostly happy with, but there are one or two lines that you don't agree with. Edit the capsule to remove or adjust those lines. Editing keeps your knowledge base accurate and aligned with your actual experience, rather than being a verbatim copy of an AI response.

Use the help icons (?) found throughout the interface for contextual guidance on specific features. You can also visit the FAQ page you're reading right now, or reach out to support for more detailed assistance.

You can reach our support team via the support form at symbol.chat/support. We aim to respond to all inquiries promptly.

We love hearing from our users! You can share feedback via the contact form at symbol.chat/support. Select "Share feedback" as the request type. Your input helps shape the future of Symbol.

If you encounter a bug, please report it via the support form at symbol.chat/support. Include as much detail as possible: what you were doing, what you expected to happen, and what actually happened. Screenshots are always helpful.

They solve different problems and complement each other. Types define knowledge templates: they tell AI how to structure and interpret stored context (e.g., "for Solutions, include symptoms, logs, and the fix"). Skills define action playbooks: they tell Claude how to perform tasks (e.g., "run tests, build, deploy"). Key advantages of types: they work with any AI tool via MCP (not just Claude Code), they produce persistent searchable capsules stored in a database, they're shareable across teams with permissions, and their guidance travels with the data so any AI reading a capsule knows how to interpret it. Use types to save and reuse knowledge; use Skills to automate repeatable workflows.