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

# Platform API Tools

> Let Ana read your workspace's dashboards, agents, playbooks, usage, and more

Platform API tools let Ana answer questions about your workspace itself, not just your data. Ana can look up dashboards, data apps, agents and their runs, playbooks and reports, datasets, ontology files, connector usage, feed posts, evaluations, roles, audit logs, and usage statistics by calling the same read-only APIs the product uses.

> [!NOTE]
>
> **Availability:** Platform API tools are a **Beta** feature. An admin must turn on **Agent platform API tools** under **Settings → Features → Beta**. Turning it on opts the whole org in.

## What You Can Ask

* "Which dashboards haven't been viewed in the last 30 days?"
* "List the runs of the Weekly Revenue agent that failed this month."
* "Which connectors were queried the most last week, and by how many people?"
* "Who changed the Sales role's permissions, and when?"
* "Show me the open ontology patches and who owns the files they touch."
* "Summarize the reports my playbooks sent this week."

## How It Works

**1. Ana picks a service**

Ana sees a list of the platform services it can read from, such as Dashboards, Agents, or Observability. When your question needs one, it first asks that service which operations it offers and what each one takes.

**2. Ana calls an operation**

Ana calls the operation that answers your question, for example listing dashboards or fetching an agent's runs. Each call shows in the thread with the service's icon and the operation's name, so you can see what was looked up.

**3. Ana works with the result**

Small responses come back to Ana directly. Large ones are loaded into the Python sandbox as a variable and a table, so Ana can filter, count, and chart them with the [Python](https://docs.textql.com/core/how-it-works/tools/python) tool instead of reading every row.

## Permissions

Turning the feature on does not grant anyone new access. In a chat, every call runs as the person chatting, with their own roles, and is checked the same way it would be in the product. If you can't see a dashboard in the UI, Ana can't read it for you either; the call fails with a permission error.

Scheduled agent and playbook runs can use these tools too, and their calls run with the permissions of the agent's or playbook's owner. Anything the owner can read can end up in that run's feed posts, Slack messages, or emails, so keep this in mind when the audience is wider than the owner.

Only read-only operations are available. Ana cannot create, edit, share, run, or delete anything through platform API tools. If you ask for a change these tools can't make, Ana will say so.

## Services

| Service | What Ana can read |
| - | - |
| Agent | Agents, their runs, and their agent databases |
| Data App | Data apps, versions, view stats, and app databases |
| Artifact | Artifacts, folders, and versions |
| Audit Log | Audit log entries |
| Thread | Threads, their history, artifacts, and LLM usage |
| Check | Checks and their latest results |
| Connector | Connectors, tables, previews, usage, and connector groups |
| Dashboard | Dashboards, folders, versions, view stats, and health |
| Dataset | Datasets, folders, values, and stats |
| Evaluation | Eval questions, runs, and summaries |
| Feed | Feed posts, comments, channels, and leaderboard |
| Observability | Activity, engagement, billing, and topic statistics |
| Ontology | Ontology files, history, patches, owners, skills, and usage |
| Playbook | Playbooks, reports, and batch runs |
| Role | Roles |
| Template | Playbook templates |

## Limitations

* Read-only. Nothing in your workspace changes as a result of a platform API call.
* Available in scheduled agent and playbook runs, where calls use the owner's permissions.
* Not available in threads limited to an API key's or OAuth client's scope, or in graded evaluation runs.
* Very large responses are summarized and loaded into Python rather than shown in full. If one is still too large, Ana will narrow the request with filters or a smaller page.


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