> ## Documentation Index
> Fetch the complete documentation index at: https://docs.oqoqo.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# oqoqo project ai and setup

> Inspect and change agents, custom agents, models, and Oqoqo AI defaults, check MCP server sign-ins and tools, and get secure setup links.

These commands set up the agents and models a project runs, and find what still needs a value in the dashboard.

| Commands | What they cover |
| - | - |
| `oqoqo project ai` | [Agents](https://docs.oqoqo.ai/setup/agents.md), [custom agents](https://docs.oqoqo.ai/setup/custom-agents.md), [model providers](https://docs.oqoqo.ai/setup/model-providers.md) and their models, and the model of each [Oqoqo AI feature](https://docs.oqoqo.ai/setup/oqoqo-ai.md) |
| `oqoqo mcp accounts`, `oqoqo setup links` | What still needs a sign-in or a secret value |
| `oqoqo mcp tools` | The tools of an MCP server |

Secret values never pass through the CLI. API keys, OAuth sign-ins, secret headers, and environment variable values go in the dashboard, and the CLI returns the exact link.

## oqoqo project ai inspect

Read agents, custom agents, provider connections, models, and the model of each Oqoqo AI feature. It never starts a model refresh or changes settings.

```bash theme={null}
oqoqo project ai inspect [flags]
```

**Plugin (MCP) tool:** [`inspect_project_ai`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#inspect_project_ai)

**Flags**

| Flag | Type | Default | Description |
| - | - | - | - |
| `--view` | `roles`, `providers`, `agents`, `models`, `agent_registry` | `roles` | What to read. See **Result fields**. |
| `--provider-connection-id` | string | | `models` view: the `aic_…` connection whose models to list. |
| `--agent-key` | string | | `models` view: the `acp-…` custom agent whose models to list. |
| `--access` | `shared`, `private` | | `providers` view: the organization's API keys, or your own accounts. |
| `--connection-state` | `connected`, `needs_attention`, `not_connected` | | `providers` view: `needs_attention` is saved but must be fixed. |
| `--availability` | `available`, `unavailable` | | `models` view of a provider: `available` lists the models that an agent or an Oqoqo AI feature can use now, and `unavailable` lists the rest. |
| `--query` | string | | `agent_registry` view: match the registry ID, name, or description. |
| `--limit` | integer | `100` | Results per page, up to 500 (default 100). |
| `--cursor` | string | | The `nextCursor` from the previous page. |

Also takes the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

**Examples**

List provider connections that you must fix:

```bash theme={null}
oqoqo project ai inspect --view providers --connection-state needs_attention
```

List a provider's models that can't be used, with the reason:

```bash theme={null}
oqoqo project ai inspect --view models --provider-connection-id aic_000000000000000000000001 --availability unavailable
```

List agents, with each custom agent's status and fix:

```bash theme={null}
oqoqo project ai inspect --view agents
```

List a custom agent's models:

```bash theme={null}
oqoqo project ai inspect --view models --agent-key acp-0123456789abcdef
```

Find an agent in the ACP Registry:

```bash theme={null}
oqoqo project ai inspect --view agent_registry --query gemini
```

**Result fields**

| View | What it lists |
| - | - |
| `roles` | Each Oqoqo AI feature (`ask` Ask Oqo, `judge` Evals, `sweep` Insights, `consolidation` Experiment insights), its saved model, and `availability`, with `availabilityReason` and `availabilityMessage` when it can't run. Also returns top-level `revision`, `organizationRevision`, `secretRevision`, and `editRevision` for writes. |
| `providers` | Each provider connection with its `access` (`shared` or `private`), `connectionState`, model counts (`modelCount`, `enabledModelCount`, `unavailableModelCount`), and model refresh state (`modelRefresh`) |
| `agents` | Every agent and its kind (`built_in` or `custom`). A built-in agent has `supportedProviders`: every provider type it works with (`templateKey`, `name`), in the order Oqoqo suggests them. The `providers` view items with the same `templateKey` say which ones work and give their `setupUrl`. A custom agent has a [`custom` block](#custom-agent-fields). |
| `models` | The models of one connection or custom agent (`--provider-connection-id` or `--agent-key`), each with its `availability` |
| `agent_registry` | The ACP Registry agents that you can create a custom agent from |

* `blockers` is empty when you can run models. Otherwise it says why not, such as no working provider, with the `setupUrl` that fixes it.
* `modelRefresh` has `status` (`queued`, `running`, `succeeded`, or `failed`), `listedAt`, and the counts of the last refresh. `listedAt` is the time the refresh read the provider's list, and `null` before that. While `status` is `running` and `listedAt` is set, the `models` view has the listed models, and the [checks](https://docs.oqoqo.ai/setup/model-providers.md#turn-off-or-refresh-models) that decide which ones you can select still run. The text output then says **Models listed; checking which ones you can use**.
* A provider's `modelCount` counts the models that an agent or an Oqoqo AI feature can use now, and `enabledModelCount` counts the ones of those that are turned on. `unavailableModelCount` counts the listed models that can't be used with the current credentials, and the text output adds **(N more can’t be used)** for them.
* Each model of a provider has `availability` (`available` or `unavailable`). An unavailable model also has `availabilityReason` (`refused`, `credentials_changed`, or `not_checked`) and `availabilityMessage`, which says the reason in a sentence. [Models that cannot be used](https://docs.oqoqo.ai/setup/model-providers.md#models-that-cannot-be-used) explains each reason.
* `--availability` works only with `--provider-connection-id`. With another view or with `--agent-key`, the command changes nothing and returns `invalid_request`. A cursor continues only the list it came from, so pass the same `--availability` with `--cursor`.
* Each model has `contextWindowTokens`, `contextWindowSource` (`catalog` or `stated`), and your `statedContextWindowTokens`. `suggestedContextWindow` (`tokens`, `catalog`, `catalogModelId`) comes from a public catalog and applies only once you state it with `set_model_context_windows`.
* A custom agent's model has `agentState`: `current`, `stale` (**Not confirmed**), or `unavailable` (**Unavailable**).

### Custom agent fields

> [!NOTE]
>
> Custom agents are in private beta. Email [support@oqoqo.ai](mailto:support@oqoqo.ai) to enable them for your organization.

In the `agents` view, each custom agent has a `custom` block:

| Field | Meaning |
| - | - |
| `status` | `needs_values`, `testing`, `needs_attention`, `not_tested`, or `ready`. The dashboard shows **Needs values**, **Testing**, **Needs attention**, **Not tested**, and **Ready**. |
| `description` | The agent's description in Oqoqo, or empty |
| `revision` | Send it as `expectedRevision` to change or remove the agent. |
| `connectionId` | The agent's own `aic_` connection. Its models are the `models` view of this ID, and `set_model_enabled` takes it. |
| `url` | The agent's page in the dashboard |
| `definition` | How the agent installs, its arguments, and its skills folder |
| `env` | Each environment variable name, whether it has a value, and the dashboard link where you set it |
| `envNote` | What to know about `env`, when there is something to know |
| `testBlockers` | Why [`test-agent`](#oqoqo-project-ai-test-agent) can't start in this project yet. Empty when it can. |
| `handshake` | The start-up details of `latestRun`, such as `failure.message`, `failure.detail`, and `authMethods` |
| `fix` | For `needs_attention`, the failed step and what to change |
| `mcpServersRefused` | `true` when a run at the current revision refused its MCP servers. Preflight then blocks treatments with MCP servers for this agent. |
| `latestRun` | The newest finished run at the current revision that wasn't cancelled and didn't refuse its MCP servers: its `runId`, `experimentId`, `status`, and `url` |

## oqoqo project ai apply

Create or change custom agents, turn agents and models on or off, state context windows, refresh model lists, and set each Oqoqo AI feature's model and instructions.

```bash theme={null}
oqoqo project ai apply --operation <json> [flags]
```

**Plugin (MCP) tool:** [`manage_project_ai`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#manage_project_ai)

**Flags**

| Flag | Type | Default | Description |
| - | - | - | - |
| `--operation` | string | | The operation as one JSON object. |
| `--operation-file` | file path | | A file that holds the operation JSON. |

Pass one of the two. `oqoqo schema project ai apply` prints the full schema. Also takes the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

### Apply operations

| `action` | What it does | Main fields |
| - | - | - |
| `set_agents` | Turn agents on or off in the project | `expectedRevision`, and either `agents` (1 to 1,000 items, `[{"key","enabled"}]`) or `enabled` for every agent (never both) |
| `set_model_enabled` | Turn one model on or off | `connectionId` (`aic_…`), `modelId` (`aim_…`), `enabled` |
| `set_models_enabled` | Turn a connection's models on or off | `connectionId`, `enabled`, and `modelIds` for a subset (1 to 5,000) |
| `set_model_context_windows` | State models' context windows on an organization connection | `connectionId`, `models` (`[{"modelId","contextWindowTokens"}]`, 1 to 5,000, context window 1 to 2,147,483,647). `null` clears one. |
| `refresh_models` | Refresh model lists in the background | `connectionIds` (1 to 200), or omit for every connection you can refresh |
| `set_roles` | Set each Oqoqo AI feature's model, effort, and custom instructions | `expectedOrganizationRevision`, `roleBindings` (each binding always includes nullable `providerConnectionId`, `providerModelId`, and `effortLevel`) |
| `create_custom_agent` | Create a custom agent | `name` (at most 120 characters), `source`, and optional `description`, `env`, and `skillsFolder` |
| `update_custom_agent` | Change a custom agent. Omitted fields keep their value. | `agentKey`, `expectedRevision`, and the fields to change |

A custom agent's `source` is `registry` (an ACP Registry agent, such as `{"kind":"registry","registryId":"gemini"}`), `manual` (an npm package, Python package, or download with its SHA-256, plus `args`), or `agent_json` (a standard `agent.json`). `env` lists variable names only; the result gives the `acp-…` key, the agent's page, and the dashboard link for each value.

**Examples**

Turn on every agent in the project:

```bash theme={null}
oqoqo project ai apply --operation '{"action":"set_agents","expectedRevision":1,"enabled":true}' --output json
```

Create a custom agent from the ACP Registry:

```bash theme={null}
oqoqo project ai apply --operation '{"action":"create_custom_agent","name":"Gemini CLI","source":{"kind":"registry","registryId":"gemini"},"env":["GEMINI_API_KEY"]}' --output json
```

Create a custom agent from an npm package with an exact version:

```bash theme={null}
oqoqo project ai apply --operation '{"action":"create_custom_agent","name":"My agent","source":{"kind":"manual","install":{"kind":"npm","package":"my-agent@1.4.2"},"args":"acp --verbose"},"env":["MY_AGENT_API_KEY"]}' --output json
```

State a model's context window:

```bash theme={null}
oqoqo project ai apply --operation '{"action":"set_model_context_windows","connectionId":"aic_4f0e2c9b","models":[{"modelId":"aim_7a13d5e0","contextWindowTokens":500000}]}'
```

Refresh model lists, then check progress:

```bash theme={null}
oqoqo project ai apply --operation '{"action":"refresh_models"}' --output json
oqoqo project ai inspect --view providers
```

**Notes**

* The provider's current context window wins over a stated one.
* After `refresh_models`, resolve any failed list's `actionRequired`, then poll `oqoqo project ai inspect` until `counts.modelRefresh.refreshing` is 0. To read a provider's models before its checks finish, read the `models` view once that provider's `modelRefresh.listedAt` is set.
* A refused operation changes nothing and exits `4` with one of these codes:

| Code | When |
| - | - |
| `validation_error` | The operation isn't valid. |
| `configuration_changed` | A revision is stale (`set_agents` or a custom agent). `details.currentRevision` has the current one. |

## oqoqo project ai test-agent

Start a Test agent run, which checks that a custom agent installs, starts, and answers.

```bash theme={null}
oqoqo project ai test-agent <agent-key> [flags]
```

**Plugin (MCP) tool:** [`test_custom_agent`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#test_custom_agent)

Uses 1 run from your balance. In Ask Oqo, you approve each call.

**Arguments**

| Argument | Description |
| - | - |
| `<agent-key>` | Required. The custom agent's key (`acp-…`). |

**Flags**

| Flag | Type | Default | Description |
| - | - | - | - |
| `--attempt-id` | string | A new ID | An ID you choose, so that a retry doesn't repeat the change. 1 to 100 characters. |

Also takes the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

**Examples**

Test an agent, watch the run, then read its status:

```bash theme={null}
oqoqo project ai test-agent acp-0123456789abcdef
oqoqo experiments watch <launch-id>
oqoqo project ai inspect --view agents
```

**Notes**

The run is one short **Agent check** task with **Agent default**, the **Raw Agent** treatment (no tools), no project files, and Evals and Insights off. Watch the launch ID with [`oqoqo experiments watch`](https://docs.oqoqo.ai/cli/experiments#oqoqo-experiments-watch). Each finished run with these settings updates the agent's status.

The command refuses while `custom.testBlockers` names a blocker, with the first one's code:

| Code | Fix |
| - | - |
| `custom_agent_test_running` | A Test agent run with these settings is in progress. Open it. |
| `custom_agent_disabled` | Turn the agent on with `set_agents` in `oqoqo project ai apply`. |
| `custom_agent_value_missing` | Set each value in the dashboard with its link. |
| `custom_agent_default_model_disabled` | Turn on **Agent default** with `set_model_enabled` and the blocker's `connectionId` and `modelId`. |

## oqoqo project ai delete

Remove a custom agent, delete or reset an organization provider connection, or disconnect your own account from a provider.

```bash theme={null}
oqoqo project ai delete --operation <json> [flags]
```

**Plugin (MCP) tool:** [`delete_project_ai`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#delete_project_ai)

> [!WARNING]
>
> Removing a custom agent clears its environment variable values. Deleting or resetting a provider connection affects every project in the organization. In a terminal, the CLI asks you to confirm; with `--output json` or no terminal, pass `--yes`.

**Flags**

| Flag | Type | Default | Description |
| - | - | - | - |
| `--operation` | string | | The operation as one JSON object. |
| `--operation-file` | file path | | A file that holds the operation JSON. |
| `--yes` | boolean | `false` | Skip interactive confirmation. Required with `--output json` or outside a terminal. |

Also takes the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

### Delete operations

| `action` | What it does | Fields |
| - | - | - |
| `remove_custom_agent` | Remove a custom agent | `agentKey`, `expectedRevision` |
| `delete_connection` | Delete a custom organization provider connection | `connectionId`, `expectedOrganizationRevision`, `expectedSecretRevision`, `expectedEditRevision` |
| `reset_connection` | Reset a built-in organization provider connection | `connectionId`, `expectedOrganizationRevision`, `expectedSecretRevision`, `expectedEditRevision` |
| `disconnect_personal_connection` | Disconnect your own account from a personal provider, such as a subscription on **Profile → Integrations** | `connectionId` |

**Examples**

Remove a custom agent:

```bash theme={null}
oqoqo project ai delete --operation '{"action":"remove_custom_agent","agentKey":"acp-0123456789abcdef","expectedRevision":2}' --yes --output json
```

**Notes**

* A removed custom agent no longer shows in new experiments or the Agents list, and its past results keep its name. Oqoqo refuses the removal while a run of the agent is in progress, either a Test agent run or an experiment. Removing it again returns the same result.
* A personal disconnect changes only your own account in this organization.
* Pass the revisions from `oqoqo project ai inspect`. A refused removal changes nothing and exits `4` with one of these codes:

| Code | When |
| - | - |
| `configuration_changed` | A stale revision on `remove_custom_agent`. `details.currentRevision` has the current one. |
| `project_ai_configuration_changed` | A stale revision on the other actions |
| `project_ai_target_not_found` | The agent or connection is already gone. |

## oqoqo mcp accounts inspect

Check whether an MCP server has the sign-in and headers it needs, without showing any secret.

```bash theme={null}
oqoqo mcp accounts inspect <mcp-id> [flags]
```

**Plugin (MCP) tool:** [`inspect_mcp_account`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#inspect_mcp_account)

**Arguments**

| Argument | Description |
| - | - |
| `<mcp-id>` | Required. The MCP server's current `mcp_` ID. |

Takes only the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

**Examples**

Check an MCP server's connection:

```bash theme={null}
oqoqo mcp accounts inspect mcp_2e8b6f13
```

**Result fields**

| Field | Meaning |
| - | - |
| `ready` | Whether the server can run. When saved values aren't [bound to the current address](https://docs.oqoqo.ai/experiments/tools.md#saved-header-values-and-the-server-address), it's `false`. |
| `recovery` | When the server can't run: a `message`, the `action` to take, `setupUrl`, `retryAt` (when to try again), and `apiKeyAlternative` when the server has an API key option |
| `setupUrl` | Where a person signs in or adds a header. For unbound saved values, it's the version's **Tools** tab, where a person selects **Refresh**. |
| `redirectUri` | For **Sign in with provider**, the [address to add to your OAuth app](https://docs.oqoqo.ai/experiments/mcp-server-sign-in.md#redirect-uri) |
| `oauthAppConfigured` | Whether a Client ID is saved |

## oqoqo mcp accounts detect-auth

Detect how an MCP server authenticates, and whether that matches its saved configuration.

```bash theme={null}
oqoqo mcp accounts detect-auth <mcp-id> [flags]
```

**Plugin (MCP) tool:** [`detect_mcp_auth_mode`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#detect_mcp_auth_mode)

**Arguments**

| Argument | Description |
| - | - |
| `<mcp-id>` | Required. The MCP server's current `mcp_` ID. |

Takes only the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

**Examples**

Detect an MCP server's sign-in method:

```bash theme={null}
oqoqo mcp accounts detect-auth mcp_2e8b6f13 --output json
```

**Result fields**

| Field | Meaning |
| - | - |
| `detectedAuthMode` | `oauth`, `headers`, `unreachable`, or `not_applicable` |
| `configurationMatches` | `false` when the detected mode differs from the saved one. Then review the server at `setupUrl`. |
| `registration` | For `oauth`: `automatic` (the provider registers Oqoqo by itself), `oauth_app_required` (you create an OAuth app and paste its Client ID and Client secret at `setupUrl`, which opens the **OAuth app** section), or `unknown`. `null` for other modes. |
| `redirectUri` | For `oauth`, the address to add to the OAuth app. `null` for other modes. |
| `oauthAppConfigured` | Whether the server has a saved Client ID |
| `setupUrl` | The dashboard page for the server |

See [MCP server sign-in](https://docs.oqoqo.ai/experiments/mcp-server-sign-in.md).

## oqoqo mcp tools list

List the tools that an MCP server listed when Oqoqo last checked it.

```bash theme={null}
oqoqo mcp tools list <mcp-id> [flags]
```

**Plugin (MCP) tool:** [`list_mcp_server_tools`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#list_mcp_server_tools)

**Arguments**

| Argument | Description |
| - | - |
| `<mcp-id>` | Required. The MCP server's current `mcp_` ID. |

**Flags**

| Flag | Type | Default | Description |
| - | - | - | - |
| `--version-id` | string | Latest version | An `mcpver_` version of the server. |
| `--query` | string | | Only tools whose name, title, or description contains this text. |
| `--after-position` | integer | | The `nextAfterPosition` from the previous page. |
| `--limit` | integer | `50` | Results per page, up to 200 (default 50). |
| `--full` | boolean | `false` | Add each tool's input schema. |

Also takes the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

**Examples**

Find the tools about issues:

```bash theme={null}
oqoqo mcp tools list mcp_2e8b6f13 --query issue --output json
```

**Result fields**

| Field | Meaning |
| - | - |
| `canTurnOff` | Whether a treatment can [turn the tool off](https://docs.oqoqo.ai/experiments/treatments.md#turn-off-mcp-tools) |
| `lastChange` | Tools the last change added, removed, or changed: up to 20 names per list, with full counts |
| `usageRunCount` | The project's runs whose treatment included this version |
| `onRunCount`, `runCount`, `callCount` | Per tool: runs with the tool on, runs that called it, and its calls. See [See how runs use each tool](https://docs.oqoqo.ai/experiments/tools.md#see-how-runs-use-each-tool). |
| `offInTreatmentCount`, `offInTreatmentIds` | Treatments that turn the tool off on any version, and up to 20 of their IDs |
| `headerValuesBound` | `false` when saved values (header values, an OAuth app's client secret) aren't [bound to the address](https://docs.oqoqo.ai/experiments/tools.md#saved-header-values-and-the-server-address). Only a person's **Refresh** at `setupUrl` binds them; until then `oqoqo mcp tools refresh` fails with `mcp_tool_check_needs_web` and launches stop with `mcp_header_values_unbound`. |
| `truncated`, `nextAfterPosition` | Whether more tools remain, and the position to pass to `--after-position` |

**Notes**

Tool titles, descriptions, and hints are the server's own text.

## oqoqo mcp tools refresh

Check an MCP server's tools now and list them.

```bash theme={null}
oqoqo mcp tools refresh <mcp-id> [flags]
```

**Plugin (MCP) tool:** [`refresh_mcp_server_tools`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#refresh_mcp_server_tools)

**Arguments**

| Argument | Description |
| - | - |
| `<mcp-id>` | Required. The MCP server's current `mcp_` ID. |

**Flags**

| Flag | Type | Default | Description |
| - | - | - | - |
| `--wait-ms` | integer | | Wait up to this many milliseconds, at most 25,000, for the check to finish. |

Also takes `--version-id`, `--query`, `--after-position`, `--limit`, and `--full`, as in [`oqoqo mcp tools list`](#oqoqo-mcp-tools-list), and the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

**Examples**

Check the server and wait for the result:

```bash theme={null}
oqoqo mcp tools refresh mcp_2e8b6f13 --wait-ms 20000 --output json
```

**Result fields**

The fields of [`oqoqo mcp tools list`](#oqoqo-mcp-tools-list) (`canTurnOff`, `lastChange`, `usageRunCount`, `onRunCount`, `runCount`, `callCount`, `offInTreatmentCount`, `offInTreatmentIds`, `headerValuesBound`, `truncated`, and `nextAfterPosition`), plus `refresh`, which is `running` until the check finishes.

**Notes**

* A check that is already running is joined, not started again. If the wait ends first, `refresh` stays `running`, and you read the list again with `oqoqo mcp tools list`.
* A server that needs your account connection or a header value returns that status and its `setupUrl`.
* The command never sends saved header values to an address they aren't bound to, and it never binds them. When `headerValuesBound` is `false`, it fails with `mcp_tool_check_needs_web` and exits `4`. Open `error.details.setupUrl`, check the address under **Refresh**, and select **Refresh**. Then run `oqoqo mcp tools list`.

## oqoqo setup links

Get the dashboard link for every secure input: account connections, API keys, secret headers, and other secret values.

```bash theme={null}
oqoqo setup links [flags]
```

**Plugin (MCP) tool:** [`get_setup_links`](https://docs.oqoqo.ai/mcp-server/agents-and-setup#get_setup_links)

Takes only the [global flags](https://docs.oqoqo.ai/cli/overview.md#global-flags).

**Examples**

Get the setup links as JSON:

```bash theme={null}
oqoqo setup links --output json
```

**Result fields**

`issues` lists each machine, MCP server, CLI, and SDK that still needs a required value or your account connection, with the `setupUrl` where a person does that step. An MCP server whose saved values aren't bound to its address has a field of scope `address`, with its **Tools** tab as `setupUrl`. The list is empty when nothing is missing.

## Related

> [card] [Custom agents](https://docs.oqoqo.ai/setup/custom-agents.md)
>
> Sources, values, status, and models of a custom agent.

> [card] [Model providers](https://docs.oqoqo.ai/setup/model-providers.md)
>
> Connect providers and state context windows in the dashboard.

> [card] [oqoqo experiments](https://docs.oqoqo.ai/cli/experiments.md)
>
> Launch the agents and models you turned on.

> [card] [Oqoqo AI defaults](https://docs.oqoqo.ai/setup/oqoqo-ai.md)
>
> Choose the model of each Oqoqo AI feature.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.