> For the complete documentation index, see [llms.txt](https://docs.kinesis.network/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.kinesis.network/mcp/openai-profile.md).

# OpenAI MCP profile

Use individual MCP actions with the OpenAI profile's credential restrictions and output schemas.

Kinesis provides an MCP profile for OpenAI integrations:

```
https://mcp.kinesis.network/mcp?profile=openai
```

It exposes **109 individual action tools** before group or read-only filtering. The normal catalog has 111 actions; this profile omits `get_byod_token`, which returns a machine provisioning credential, and `redeem_voucher`. Redeem voucher codes under Billing in the portal. It uses the same authentication, scopes, organization roles and access restrictions as other MCP connections.

## How it differs

| Behavior                                             | Default compact connection                      | `profile=openai`                                            |
| ---------------------------------------------------- | ----------------------------------------------- | ----------------------------------------------------------- |
| Calls                                                | Group tool with `{action, params}`              | Action name with its arguments directly                     |
| Catalog                                              | 18 tools covering 111 actions                   | 109 individual tools                                        |
| `describe_action`                                    | Available for loaded folded groups              | Not present; inspect the action's advertised schema         |
| Result schemas                                       | Full response schemas in this reference         | Advertised `outputSchema` describes top-level result fields |
| Machine provisioning token                           | `get_byod_token` available with its permissions | Omitted; attach your own machine through the portal         |
| Voucher redemption                                   | `redeem_voucher` available with its permissions | Omitted; redeem in portal Billing                           |
| Private registry login and secret environment values | Supported where the action schema allows        | Configure in the portal                                     |

Leave `mode` out of this URL, or use `mode=all`. `mode=compact` with this profile is rejected as `400 invalid_request`. Group selection and `readonly=true` still work, for example:

```
https://mcp.kinesis.network/mcp?profile=openai&groups=account,projects,apps&readonly=true
```

`read_docs` remains included. Include `account` when you need `get_current_user` or `list_orgs`.

## Call actions directly

Call `list_projects` with:

```json
{"org":"acme","limit":50}
```

Call `create_app` directly with its input fields, not an `action` / `params` wrapper. Organization-level calls take `org`; project, app and grid calls resolve the organization from the resource. Read and write annotations remain attached to each individual action.

## Handle credentials in the portal

This profile removes `registryAuth` and environment-variable `secret` fields from advertised input schemas. Runtime checks refuse a supplied private registry login and nonempty environment values that are explicitly secret, already stored as secret, or detected as credentials with high confidence. These refusals are validation errors; switching tool names does not bypass the rule.

Set those values in the Kinesis dashboard. An agent should explain where to set them instead of asking the user to paste credentials into the conversation. Ordinary non-secret configuration can still be supplied through tools.

When reading app environment variables, secret values are null as usual. This profile also returns null for values that look like credentials even if they were not marked secret. Send null back on update to preserve the stored value. A new variable still needs a value; null cannot create one.

`envFrom` can copy an accessible app's environment on the server without returning its secrets. Credential restrictions also apply to values you supply as overrides. Preserve the existing access and scope requirements.

## Logs remain raw output

The credential handling above is not a general log-redaction service. App logs and log exports contain what the app printed; build logs contain what its build steps printed; server logs contain what the machine wrote. Those outputs may contain sensitive data. Request only the logs needed for the task and avoid repeating secrets in responses.

Successful calls return JSON in text content and `structuredContent`. An advertised output schema describes top-level fields, not every nested field; the reference pages provide complete service response schemas. Error handling, safe retries and explicit `confirm: true` arguments follow the normal [agent workflows](https://docs.kinesis.network/mcp/workflows.md).

This connection profile describes the API contract; it does not imply a plugin-directory listing or approval status.

[Connect an MCP client](https://docs.kinesis.network/mcp/connect.md) · [Tool reference](https://docs.kinesis.network/mcp/tools.md)


---

# 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.kinesis.network/mcp/openai-profile.md?ask=<question>&goal=<user_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.
