Context Layer for AI assistants
Ask Claude about a competitor. Get your team's cited answer.
Connect KeystoneIQ once and your assistant reads your cited briefs, current battlecards, verified facts, and approved messaging straight from your workspace. Read-only. Answers carry their sources.
MCP is the open standard assistants use to read outside data. KeystoneIQ is one server you add once, in Claude Code, Claude Desktop, Cursor, Windsurf, or any MCP client. Starter reads briefs and verified facts (300 calls a month). Growth adds battlecards, threat scores, deal context, and approved messaging (3,000 calls a month). Pro is unmetered.
What your assistant can read
Cited briefs
Ask what changed with a competitor this week and get the brief your team already trusts, with citations.
Current battlecards
Pull the live battlecard into a deal review or a deck, with your team's approved counters for that competitor. It carries the freshness date and the sources. Growth and Pro.
Verified facts
One current value per attribute: pricing, headcount, funding, ratings. Sources that disagree are flagged, never averaged.
Approved messaging
Your team's approved positioning: who it's for, why you win, how to talk about you against each competitor. Every claim is labelled backed or team's call; drafts are never served, so the assistant only says what you stand behind. Growth and Pro.
Your team's rules
Tell KIQ what never to do, such as compete on price or promise roadmap items, and your take on the market. Briefs, search results, facts, threat scores, battlecards, deal context and approved messaging reach your assistant with those rules attached as instructions, so it follows your stance when it answers from them. Editing your stance needs Starter or above.
Why written rules matter for AI messaging →Connect in minutes
- Create an API key under Developer in KeystoneIQ.
- Add
https://keystoneiq.ai/api/mcpas an MCP server in your assistant with the key as a bearer token. Copy-ready snippets for each client are in the connector docs. - Ask: “What changed at [competitor] this week, and what should our rep say on a call?”
Try it before you sign up
Add https://keystoneiq.ai/api/mcp with no key and your assistant answers from a sample workspace. It answers as Kestrel, a sample project-management company that tracks Asana, Monday.com, ClickUp, Notion, and Linear. Ask it: “What changed at Asana this week, and what should our rep say?”
New to MCP? Read what an MCP context layer does for a PMM, including setup and what to look for in any vendor's server.
Tools your plan does not include still answer, with a clear upgrade note instead of an error, so the assistant can tell you what to do next.
MCP questions
- What is the KeystoneIQ MCP server?
- It is a read-only Model Context Protocol server at https://keystoneiq.ai/api/mcp. Add it once to Claude Code, Claude Desktop, Cursor, Windsurf, or another MCP client that supports streamable HTTP, and your assistant reads your competitive intelligence straight from your KeystoneIQ workspace. Browser-hosted assistants that need OAuth sign-in (claude.ai on the web, ChatGPT) are not supported yet.
- Which plans include it?
- Starter reads competitors, search, briefs and verified facts, with 300 tool calls a month. Growth adds battlecards, threat scores, deal context and approved messaging, with 3,000 calls a month. Pro is unmetered. The Free plan does not include MCP; the 14-day trial does.
- Does my assistant follow my team's rules?
- Yes, once your team sets its stance. Eight content tools (search, latest brief, brief, battlecard, threat scores, facts, approved messaging and deal context) return your rules as team_rules, with an instruction to follow them when paraphrasing and never to do anything listed under never. Editing the stance needs Starter or above.
- Can my assistant change anything in KeystoneIQ?
- No. Every tool is read-only, every call is scoped to the workspace tied to your API key, and every call is recorded in your workspace's API audit log.
- Can I try it before I sign up?
- Yes. Add https://keystoneiq.ai/api/mcp with no key and your assistant answers from a sample workspace for a fictional company, Kestrel. It is limited per IP address and has a shared daily allowance.
Give your assistant context it can cite.
Start your 14-day trial, connect in minutes, and ask your first competitive question from the tool you already work in.