Docs Connect

Use fixvis from Claude Code, Codex or ChatGPT

The fixvis MCP server lets your own AI assistant read your issues, fix them in your code, move them on the board, verify them and draft content. Available on Starter and above.

Server address: https://api.fixvis.ai/mcp

Claude Code

Add the server

claude mcp add --transport http fixvis https://api.fixvis.ai/mcp

Sign in
Run /mcp, choose fixvis and select Authenticate.
Approve access
In the browser, sign in, pick the organisation, tick the projects this connection can use, choose an access level and select Approve.
Check it
/mcp shows fixvis as connected. There is no token to copy.

Codex

Add the server

codex mcp add fixvis --url https://api.fixvis.ai/mcp

Sign in
Complete the same browser sign-in and approval.
```
codex mcp login fixvis
```
Check it
```
codex mcp list
```

ChatGPT and other MCP clients

In ChatGPT, turn on Developer mode in settings, then add fixvis as a connector with the server address above. Any other client that supports remote MCP servers over HTTP, such as Cursor, connects to the same address.

Access levels

Level	What the assistant can do
Read your fixvis data	Read-only: issues, markets, competitors, content gaps, outcomes
Read and update	Also move issues, verify fixes, create content tasks and record published URLs. Scans and content generation still ask you to confirm

Tools

Tool	Access	What it does
`list_projects`	Read	Projects in this connection, with latest audit and open issue counts
`list_issues`	Read	A project's issues, worst first, filtered by area, severity and status
`get_issue`	Read	Everything needed to fix one issue: evidence, affected URLs, how it's re-checked
`explain`	Read	An issue or term in plain language, and why it matters
`move_issue`	Write, free	Move an issue to To do or In progress, or mark it shipped
`verify_fix`	Write, free	Re-check the live page and return the outcome with evidence
`list_markets`	Read	Tracked markets, the primary one, and each one's latest scan
`get_market_visibility`	Read	One market's AI visibility score, trend, citations, issues, and the questions you win or lose
`get_markets_rollup`	Read	All markets side by side
`list_competitors`	Read	Tracked competitors in one market and where you stand
`list_content_gaps`	Read	Content opportunities in one market, ranked
`get_content_gap`	Read	One opportunity in full: keywords, competitor URLs and a brief
`create_content_task`	Write, free	Add a content piece to the board, from a gap or a topic
`generate_content`	Write, uses a fix	Write a brief or draft for a content task
`get_content`	Read	A generated brief or draft, or its status
`set_published_url`	Write, free	Record the live URL of a published piece so fixvis can measure it
`get_outcomes`	Read	What fixes and content achieved over a period
`list_delivered_fixes`	Read	Fixes and content delivered in a period, with outcomes
`get_fix_outcome`	Read	One delivered fix's before and after
`run_market_scan`	Write, uses a fix	Scan one market or all markets

Anything that uses a fix asks first

generate_content and run_market_scan work in two steps. The first call returns a preview of the cost and your balance, and nothing is spent. Only when you agree does the assistant call again to run it. A failed draft is not charged. Today's included scan of each market is free.

Things to ask

“Use fixvis to show the open technical issues for gorusmeyap.com, worst first.”
“Get the full details of the top issue and find where in this repo it comes from. Don't change code yet.”
“Apply the fix, then move that issue to In progress in fixvis.”
“I've deployed. Verify that fix with fixvis.”
“What should we write next for Germany?”
“How did last month's fixes do?”