> For the complete documentation index, see [llms.txt](https://docs.feathery.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.feathery.io/develop/mcp-server/security-limits-and-troubleshooting.md).
# Security, Limits and Troubleshooting
How Feathery MCP access works, how to keep API keys safe, and how to fix common connection issues.
### How Access Works
| | Sign in with Feathery | API key |
| --------------------- | ------------------------------------------------------ | ----------------------------------- |
| **Acts as** | You | The API key's environment |
| **Data it can reach** | The organization you choose, in its live environment | The key's environment, live or test |
| **What it can do** | What you allowed when connecting, limited by your role | Everything in that environment |
| **Stays connected** | Up to 90 days, or until unused for 30 days | Until the key is rotated |
| **To disconnect** | Remove Feathery from your AI tool | Rotate the API key |
Some tools also depend on your plan and features. Workspace tools need a plan that includes the developer API, AI document extraction tools need Document Intelligence, and meeting tools need the AI Notetaker.
### Keep Your API Key Safe
* Store the key in your AI tool's secret or environment variable settings when possible.
* Don't commit API keys to shared configuration files.
* Use a test environment API key when experimenting.
* Rotate the API key if it is shared accidentally.
### Review Actions That Change Data
Some tools create, update, or delete Feathery data; [Supported Tools](/develop/mcp-server/supported-tools.md) marks each one. Turn on tool approval in your AI tool so you can review these actions before they run. Deletion and removal tools also need an explicit confirmation before they run.
### Troubleshooting
#### Connecting
Feathery doesn't appear in my AI tool
Confirm the MCP server was saved with the right URL for your region, then restart or reload your AI tool.
Authentication fails
If you signed in, remove Feathery from your AI tool and connect again. If you use an API key, confirm the key is valid and that the header value starts with `Token`.
The sign-in page says the request expired
Connection requests last 10 minutes. Start connecting again from your AI tool.
The sign-in page says nothing is available
Your organization's plan doesn't include the developer API, or your organization doesn't use the AI Notetaker.
The sign-in page says to connect from your own login
You opened it from an embedded workspace. Log in to Feathery directly and connect again.
The assistant is connected to the wrong organization
Each connection covers one organization. Remove Feathery from your AI tool, connect again, and choose the other organization.
#### Using the Tools
The assistant can't find expected forms or submissions
Confirm you connected to the right organization. With an API key, confirm the key belongs to the right environment.
An action isn't allowed
Your Feathery role doesn't allow it. Ask an Admin to run it, or to change your role.
An AI document extraction fails
Confirm Document Intelligence is enabled for your organization and that the file is accessible. If a public URL doesn't work, download the file and attach it through your AI tool instead.
The assistant is rate limited
Ask it to retry more slowly or to narrow the request.
### Frequently Asked Questions
Can I use an SDK key?
No. Sign in with your Feathery account, or use an environment API key from the **Developer** tab in your Feathery settings.
Can I connect more than one organization?
Each connection covers one organization. To switch, remove Feathery from your AI tool, connect again, and choose the other organization.
Why are some Feathery features missing?
Feathery MCP supports a focused set of tools. More Feathery actions may be added over time.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.
Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:
```
GET https://docs.feathery.io/develop/mcp-server/security-limits-and-troubleshooting.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.