Every session starts from zero.
Your AI assistant is brilliant today. Tomorrow it won't remember why you chose Postgres, that the auth system is being refactored, or the edge case that took you hours to debug.
You explain. Again.
Memory that persists.
Cog gives your AI a knowledge graph. Store decisions, recall context, connect ideas. Knowledge strengthens over time.
cog_remember
Store knowledge
Capture decisions, context, and insights as your AI works.
cog_remember({
term: "Database Choice",
definition: "PostgreSQL for JSONB support"
})
cog_recall
Search by meaning
Semantic search finds relevant knowledge, not just keywords.
cog_recall({
query: "data storage approach"
})
// → finds PostgreSQL decision
cog_associate
Connect ideas
Build a knowledge graph where connections strengthen over time.
cog_associate({
source: "Auth System",
target: "JWT Tokens",
predicate: "uses"
})
One config.
Cog speaks MCP. Add it to Claude Code, Cursor, or any MCP-compatible assistant.
{
"mcpServers": {
"cog": {
"url": "https://trycog.ai/mcp",
"headers": {
"Authorization": "Bearer cog_..."
}
}
}
}
That's it.