Integration · Direct your agents
Grok checks the work. You make the call.
Add Annsa to Grok and it reviews against the spec your agents wrote, then files what it finds with a severity you can overrule.
What it does for you
Same list, same spec as Claude Code and Cursor.
Choose what to build.
Its findings land in the same ranked list as customer feedback, with a severity (P0 broken to P3 minor) you can overrule.
Direct your agents.
It reviews against the spec's Done When lines, not a guess at what you meant.
Learn as you go.
What it finds on shipped work comes back as a finding, beside your customers' feedback. A person makes the call.
The proof
Connect, then ask.
Run the line, approve the sign-in, and ask for your priorities. The first reply tells you which workspace it is talking to.
grok mcp add --transport http annsa https://app.annsa.ai/mcpHow it connects
One line in the terminal.
Add the server, approve the sign-in, then ask for your priorities.
Add the server.
`grok mcp add --transport http annsa https://app.annsa.ai/mcp`, from xAI’s Grok Build MCP guide.
Approve the sign-in.
Your browser opens to Annsa’s consent screen. Approve it once.
Check it worked.
Ask Grok for your priorities. The reply’s `who` should say production, app.annsa.ai and your workspace.
Try this first.
“Show me the top priority and why it ranks there.” Then: “Review this branch against #N’s Done When lines and file what you find.”