Troubleshooting

The tools do not appear

This is the common one, and the cause is almost always the key.

A key the platform does not accept fails the whole connection. Your client will report that it could not connect, rather than connecting with fewer tools. So "no tools" and "cannot connect" are usually the same problem wearing two faces.

Check, in this order:

1. The key itself. Copy it again from your workspace settings. It has to start with mk_live_ and contain a dot. A key that was cut short when pasted is the single most common cause.
2. The header. In Cursor, "Authorization": "Bearer YOUR_KEY" — the word Bearer, one space, then the key. "X-Api-Key": "YOUR_KEY" works too, without a prefix.
3. A placeholder that was never replaced. If the configuration still holds something like ${QAMERA_API_KEY} or YOUR_QAMERA_AI_API_KEY, that literal text is what gets sent.
4. Stray whitespace or a line break inside the key.
5. Whether the key was revoked or has expired. Check it in workspace settings and create a new one if needed.
6. Only then, the client. Restart it after saving the configuration file, and confirm the JSON is valid — no trailing commas, and the server entry under mcpServers.

The server says your key could not be checked

A different message, and a different problem: it means the platform could not be reached to confirm the key, not that the key was rejected.

Do not create a new key. Yours was never refused. Wait a moment and reconnect. If it persists, check the server first:

curl https://mcp.qamera.ai/health

and then contact us — this one is on our side, not yours.

Everything works except checking generation status

If your assistant can list products, order generation and show previews, but reports an error whenever it checks whether a generation finished, the key is missing one permission.

Status checks need Read Content Status (content.read). Everything else needs Register Content (content.register). Nothing is pre-selected when you create a key, so a key with only the first is easy to end up with.

This looks like a stuck generation, and it is not. The generation is running normally; only the question about it is being refused. Create a new key with both permissions selected, update your client's configuration, and delete the old key.

Cannot connect at all

If the connection fails before any key question arises:

1. Run the health check above. If it answers, the server is up and the problem is closer to you.
2. Check that the address in your configuration is exactly https://mcp.qamera.ai — with https, and with no path after it.
3. Check a firewall or proxy. If you are still on the older address, it uses port 3002, which some corporate networks block; the current address uses ordinary HTTPS.
4. If you are on a VPN, try without it.

Requests start failing intermittently

You are being rate limited. Ask your assistant to slow down — put a pause between batch operations, and avoid running several sessions against the same key at once. If your work genuinely needs a higher limit, get in touch.

Generation fails with a credits error

Check your balance under Credits in Qamera AI and top up or change plan if needed.

Reading lists, opening previews and recording approvals cost nothing. Only the steps that generate — packshots, photo shoots, models, sceneries, videos — spend credits.

Still stuck

Contact the Qamera AI team through your usual support channel, and include what your client reported and roughly when. Never send your API key, in a message or a screenshot; if you think it has been exposed, delete it in workspace settings and create a new one.