KnitKnot
You are reading the agent-optimized layer of this page: the literal markdown we serve to AI crawlers and assistants, shipped in the page source of every visit. Making sure AI reads the right facts about a company is literally what KnitKnot does.

# KnitKnot MCP Server and Workflow Integrations

Bring your AI presence evidence into the assistants and team systems where work already happens.

## One workspace, multiple surfaces

KnitKnot MCP exposes a curated set of workspace tools over the same data used by the console. Customer-facing tools cover competitive position, score trends, competitors, issues and their evidence, playbooks, source intelligence, demand topics, prompts, prompt coverage, and workspace context. A playbook status can be updated from the connected assistant.

## MCP security and scope

OAuth connections use the signed-in KnitKnot identity. API keys support programmatic clients. Requests are scoped to the caller's workspace, and customer tool discovery is controlled by an explicit allowlist.

## Notion

KnitKnot creates a playbook database in the selected Notion destination. Playbooks sync into Notion, while supported status and assignment changes flow back to KnitKnot.

## GitBook

KnitKnot can validate a GitBook API token and connect the workspace to a selected GitBook space.

## Google Analytics

KnitKnot pulls observed AI-referred sessions and landing pages from the selected GA4 property and places them alongside the AI Presence Score trend. Referral analytics are directional and undercount total AI influence.

## Related resources

Raw mirror of this content: https://knitknot.ai/product/mcp-integrations.md. Site-wide summary: /llms.txt · full content: /llms-full.txt

MCP & integrations

Put your AI presence inside the tools where work happens.

Query evidence and work playbooks through KnitKnot MCP, coordinate the backlog in Notion, connect observed traffic from Google Analytics, and configure a GitBook destination. 

list_issues
1. AI recommends Northstar over you on access controls
High priority | 14 affected answers | evidence attached
get_issue -> get_playbook
Revise /security/access-controls
Includes losing quotes, cited pages, buyer vocabulary, and target terms
update_playbook_status -> in_progress
One source of truth

Your assistant reads the same evidence as the console.

MCP is a programmatic door into the KnitKnot workspace, not a separate dataset. Ask for the competitive picture, drill into an issue, inspect its playbook, or move the work forward without losing the evidence chain.

Diagnose

Read competitive position and score trends, inspect competitors, rank persistent issues, and retrieve the exact evidence behind a finding.

Act

Explore demand topics and prompt coverage, open an evidence-grounded playbook, review source intelligence, and update work status.

Connected workflows

Move the evidence into execution.

Each integration has a narrow job: coordinate the backlog, configure a documentation destination, or add an observed traffic signal to measurement.

Notion

Work the playbook backlog with your team

Sync playbooks into a Notion database and carry supported status and assignment changes back into KnitKnot.

GitBook

Configure the documentation destination

Validate a GitBook API token and select the connected space as the configured documentation destination.

Google Analytics

Place observed AI traffic beside the trend

Pull AI-referred sessions and landing pages from GA4. This is a directional referral signal, not a complete measure of AI influence.

Access

Connect with an account or an API key.

Interactive clients can use OAuth 2.1 with the same KnitKnot identity and workspace access as the console. Scripts and non-interactive clients can use a KnitKnot API key. Customer-visible MCP tools are controlled by an explicit allowlist.

OAuth 2.1

Approve an interactive connection with the signed-in KnitKnot account.

API keys

Authenticate programmatic clients without an interactive browser flow.

Workspace scope

Tool calls resolve against the caller's authorized KnitKnot workspace.

Choose the surface

Use MCP for flow. Use the console for depth.

MCP is useful when a question or status change belongs inside an assistant workflow. The console remains the richer surface for scanning tables, comparing periods, editing reports, and reviewing dense evidence.

Reach for MCP

  • Ask which competitor or issue needs attention.
  • Bring source evidence into a drafting workflow.
  • Review prompt coverage while planning research.
  • Move a playbook into progress from an assistant.

Reach for the console

  • Scan the full issue, source, or prompt catalog.
  • Compare trends and measurement periods visually.
  • Review claim-level evidence in context.
  • Configure integrations and publish reports.
Read the MCP setup guide Explore the Measurement Loop
Questions

MCP and integration questions

Start with a benchmark

Bring the benchmark into your workflow.

Start with the evidence, then connect the tools your team uses to act on it.