iFixAi runs an operational-misalignment diagnostic against an AI agent you already have deployed. This connector lets Claude or ChatGPT drive it: describe your agent, register its endpoint, run inspections, and read what went wrong.
iFixAi dials your agent over its own HTTP endpoint and puts it through up to 102 inspections: does it invent facts, exceed its authority, leak data it shouldn't, change its story under pressure, act when it should escalate. You get a graded report with the exact prompt that caused each failure and the agent's exact reply.
It reports what your agent got wrong. It does not prescribe fixes.
run-inspection hands you a command that runs them on your machine with your own LLM key. Nothing is billed and nothing is saved to iFixAi.request-credits.test-connection before spending anything. Free local runs have no such limit: your machine dials the endpoint, localhost included.Add the server URL in your client's connector settings:
https://mcp.ifixai.ai/mcpYour client will open a browser window to sign in to iFixAi and ask you to approve access. That's standard OAuth 2.1. The connector never sees your password, and iFixAi never sees your agent's credentials after you store them.
Not sure which plan you're on? Ask for get-plan first. Every paid-only tool tells a free caller plainly what it does and that a paid workspace unlocks it, rather than leaving you guessing.
On the free plan, steps 2–3 fall away: run-inspection returns the local command directly, with your endpoint and fixture file passed as arguments.
Why the fixture matters. Without one, iFixAi grades your agent against a generic placeholder, and the result is meaningless. The connector refuses to start a run that would do that.
Six tools answer on the free plan: get-plan, list-inspections, run-inspection, request-credits, search, fetch. The rest need a paid workspace and say so plainly when they don't have one.
| Tool | Does | Costs credits |
|---|---|---|
| get-plan | Which plan this account is on and where runs execute | no |
| get-wallet | Credit balance for your workspace | no |
| list-inspections | The inspection roster and what each one checks | no |
| author-fixture | Turns your agent's description into a fixture | no |
| validate-fixture | Which inspections a fixture can run, and why the rest can't | no |
| save-fixture | Stores a fixture against a connection, so a rerun needs no re-authoring | no |
| get-fixture | The fixture saved for a connection | no |
| estimate-run-cost | Credit quote plus coverage warnings | no |
| create-connection | Registers your agent's endpoint and credential | no |
| test-connection | Dials your agent once to check it answers | no |
| list-connections | Endpoints registered in your workspace | no |
| run-inspection | Paid: starts a billed hosted run, returns a run id immediately. Free: returns the command to run locally | paid only |
| get-run | Status, grade, score, credits quoted and settled | no |
| list-runs | Your runs, newest first | no |
| get-deliverable | The frozen report for a finished run | no |
| get-coverage | Which inspections your recent runs already exercised | no |
| request-credits | Raises a credit request, priced from your billing term | no |
| list-credit-requests | Your credit requests and where each one stands | no |
| cancel-run | Stops a run; you pay only what it already used (in 50-credit blocks, never more than the quote), the rest of the hold returns | part refund |
| search | Find inspections and your runs by keyword | no |
| fetch | One search result as a document | no |
Suites range from smoke (3 inspections) to all (102 hosted; the free roster is the 45 open-source inspections). Quote first: estimate-run-cost tells you two things that aren't obvious from the suite name.
all returns a capped grade. Smaller suites are useful diagnostics, not verdicts.| pass | The agent met the inspection's threshold. |
| fail | It didn't. The report carries the prompt and the reply. |
| inconclusive | The inspection couldn't reach a verdict, usually a surface your endpoint doesn't expose. Not a failure, and not a pass either. |
| Message | Means |
|---|---|
| this needs a paid workspace | You're on the free plan. The message says what the tool does and what paying adds; request-credits upgrades. |
| workspace has no tenant | Provisioning didn't finish. Ask us. |
| not enough credits | Top up before running. |
| this run has no fixture | Author one first, see step 1. |
| backend unreachable | Ours, not yours. Try again shortly. |
A run keeps going even if your client disconnects. Poll get-run with the id to pick it back up, you're never billed twice for the same start.
Email support@ifixai.ai. Include your run id if it's about a specific run.