Keep the project memory
Models come and go. Your project memory stays.
Lyriks keeps a permanent memory of your project: spec, rules, decisions and their sources, on your servers, readable by whichever LLM you choose.
Sound familiar?
- Nobody remembers why a rule was decided.
- Each new LLM means reconfiguring it all.
- Decisions are scattered across chats.
- Knowledge everywhereIn chats, tickets and heads. Scattered, then lost.
- Lyriks keeps itRules, decisions and their sources, in one model, on your servers.
- Plug in any LLMUnplug one model, plug in another. The memory stays.
What Lyriks keeps
- 01
One knowledge base
Code, Jira, Confluence, Notion, Figma and GitHub brought together.
- 02
Full traceability
Every rule linked to its source, its decision and the code that implements it.
- 03
Any LLM, through MCP
Claude Code, Cursor, Copilot or a local model read the same memory.
- 04
On your side
Installed on your servers, exportable in full. No lock-in.
What you get
- The history of every decision, with its source
- Rules and a spec that outlive team changes
- Freedom to switch LLM at any time
- An audit trail for review and compliance
- An exportable memory, with its traceability
Further reading
Models Change. Memory Stays.
Coding assistants turn over every few months. What you know about your software, and why it works that way, has to stay with you, whatever the model.
Running an AI Tool Without Your Code Leaving the Building
A hospital IT team wants AI help on its appointment module. The CISO has three questions. Here is how an on-premise, air-gapped setup answers each one.
Claude Code, Cursor, Copilot: give your agents the right spec with MCP
A coding agent builds from the context it gets. Over MCP, it can work from a validated spec with rules and sources, then be checked rule by rule.
Other use cases