> ## Documentation Index > Fetch the complete documentation index at: https://deepline.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Connect Attio to Deepline > Connect your Attio workspace, build a Play, and write researched company and contact data back to your records. [Human guide: work with your agent →](https://deepline.com/docs/human/providers/attio/guide.md) > [!NOTE] > > Connect the intended Attio workspace and verify one authored read or approved > update against its real record. Deepline connects your Attio workspace to Plays: reusable workflows that read records, research or enrich them, and write the results back to Attio. You choose the records, enrichment sources, and attributes each Play uses. ## Set up the integration 1. [Sign in to Deepline](https://code.deepline.com/sign-in) or create an account, then select the Deepline workspace that should connect to Attio. 2. Follow the public [Deepline installation guide](https://code.deepline.com/INSTALL.md) to install the CLI and complete setup. Deepline supports Claude Code, Codex, Cursor, Gemini CLI, and Antigravity. You can give your agent this instruction: ```text theme={"languages":{"custom":["/languages/deepline-prompt.json"]}} Install and set up the official Deepline CLI globally from npm (package: deepline) by following https://code.deepline.com/INSTALL.md. ``` 3. Open [Integrations](https://code.deepline.com/dashboard/integrations), search for **Attio**, and select **Connect** on its card. 4. On Attio's authorization screen, select the Attio workspace you want to use, review the requested permissions, and authorize Deepline. You need permission to install apps in that Attio workspace. 5. Return to Deepline. Confirm that the Attio card shows a connected account. Use **Reconnect** if an existing authorization needs to be refreshed. 6. Ask your agent to create a Play, check it, and run it against one record before expanding to a larger batch. Open that record in Attio to verify the result. > [!NOTE] > > The selected Deepline workspace owns the connection. Check your workspace > before connecting or running a Play. An Attio connection in another Deepline > workspace does not automatically grant access to this one. ## Read and update Attio records Plays can query people, companies, deals, users, workspaces, and custom objects. They can read record IDs and attribute values, filter records, and paginate results. For example, a company Play can read `name` and `domains`, research the company's website, and update `description` or a custom research attribute. A people Play can use names, email addresses, company relationships, and job titles as inputs, then write the verified results to writable attributes. Deepline does not write every attribute by default. The Play explicitly maps each result to an Attio attribute slug or ID. Inspect the object's attributes first: Attio controls which attributes are writable, their types, and the value formats they accept. Read-only system attributes cannot be updated. Use filtered queries to choose an existing record, and assert operations when your workflow needs an upsert. Creating a record and updating a record are separate operations. Decide how your Play should handle an existing value before overwriting it. ## Create and trigger a Play Describe the workflow to your agent, including the Attio object, selection filter, research source, destination attributes, and overwrite rule. For example: ```text theme={"languages":{"custom":["/languages/deepline-prompt.json"]}} Create a Deepline Play that reads one selected company from Attio, researches its public website, writes a company summary and source URL to custom Attio attributes, and reads the record back to verify the values. Show me the mapping before running it on more records. ``` Check the authored Play with `deepline plays check your-play.play.ts`. Run it on demand through the CLI or API, or publish it and run it from the Deepline dashboard. Plays can also run on a configured cron schedule or receive events through a Play webhook. An event-driven Attio workflow requires explicitly configuring the Attio subscription and the authenticated Play webhook. connecting Attio alone does not create a trigger. Managed Attio Monitors are not currently offered for new setup. The dashboard shows the Play's flow, run status, and results. After a write step, read the record back through the integration and open it in Attio to verify the saved values. ## Work with lists and workspace context Plays can read lists, add or update entries, and map results to list attributes. A Play that needs a new destination can create a list, object, or custom attribute. These configuration changes happen only when that operation is included in the workflow. Workspace-member access is read-only. It lets a Play resolve member identities for tasks and ownership references. Deepline does not need permission to add, remove, or change workspace members. ## Notes, tasks, meetings, and call recordings When included in a Play, Deepline can read notes and comments as research context, create notes containing a summary, and create or update tasks for follow-up. Comments and threads can be read, and a Play can post a comment. Meeting and call-recording access is read-only. A Play can read meeting details, available recording metadata, transcripts, and speakers to prepare an account brief or meeting summary, then write the output to a note or record. Deepline does not request permission to edit meetings or call recordings. These operations are not automatic: the authored Play chooses which context to read and where to write its output. Recording availability depends on the connected Attio workspace and the access Attio grants. ## Permissions The full integration supports the following features. A record-only enrichment Play uses record access and object-configuration reads. The additional scopes support the explicitly authored operations described below. | Scope | Feature that uses it | | - | - | | `record_permission:read-write` | Query, create, assert, and update records and their writable attributes. | | `object_configuration:read-write` | Inspect object and attribute schemas. create or update objects and custom attributes when a Play needs a destination. | | `list_entry:read-write` | Read, add, and update entries and their attribute values. | | `list_configuration:read-write` | Inspect lists and their attributes. create or update lists and destination attributes. | | `note:read-write` | Read research context and write notes or summaries on records. | | `task:read-write` | Read, create, and update follow-up tasks. | | `comment:read-write` | Read comments and threads and post comments from a Play. | | `user_management:read` | Resolve workspace members for task assignments and ownership references. | | `webhook:read-write` | Read and manage Attio event subscriptions for explicitly configured event-driven workflows. | | `meeting:read` | Read meeting context for briefs and summaries. | | `call_recording:read` | Read available recording metadata, transcripts, and speakers for research and summaries. | ## Verify and troubleshoot * **Attio is not connected:** select the correct Deepline workspace, return to Integrations, and connect or reconnect Attio. * **Authorization is denied:** confirm you can install apps in the selected Attio workspace. If Attio reports an app installation limit, contact Deepline support. repeatedly reconnecting will not bypass that limit. * **An update fails:** inspect the destination attribute's type and writable status, then compare the payload to Attio's accepted value format. * **A result is missing:** inspect the Play run, verify the selected record ID, and read the record back. A completed research step does not itself prove that a later write succeeded. Contact [team@deepline.com](mailto:team@deepline.com) for help with the integration. ## Verify this task Inspect the object, record ID, attribute type, and current value. Verify the destination after an approved update and preserve uncertain outcomes. Ask your agent to explain a failed check before repeating work or expanding the scope. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.