Skip to main content
Human guide: work with your agent →
Select an interface, authorize one workspace, and inspect a candidate workflow before executing it.
Use the interface that matches where your agent runs. Each option works with the same Deepline Plays, Monitors, Database, integrations, and run history.

Claude Code

Install the Deepline plugin and Skills when Claude Code is where you work.

Claude Cowork

Use the Deepline plugin from Claude Desktop with Cowork.

MCP

Connect Deepline to ChatGPT Work, Claude Desktop, or another MCP client.

Grok Bot

Run Deepline from xAI’s desktop assistant using its computer access and skills.

Hermes

Run Deepline through the Hermes compatibility pack when Hermes is your operator interface.

Build a custom agent

Start from the open-source Deepline GTM agent reference implementation for a Slack, REST, or web-chat agent backed by the Deepline API.

​
Choose the right path

The custom-agent reference keeps Deepline responsible for provider routing, integrations, billing, Play execution, and run history. Your agent owns its interface, authentication, and user experience.

​
Run a job with a Deepline skill

After installing through Quickstart, use deepline-gtm for a GTM request or choose a specialized entry from the skill catalog.
  1. Define the output, format, evidence, and acceptance checks.
  2. Inspect the input and preserve existing good data.
  3. Search and describe the tools or Plays that may execute the job.
  4. Run a small pilot.
  5. Review coverage, evidence, failures, and the Deepline-facing spend estimate.
  6. Approve the full run.
  7. Export the result or save the accepted workflow as a reusable Play.
For durable code, schedules, and webhooks, use deepline-gtm with the Play authoring guide and deepline plays --help.

​
Verify this task

Verify authorization in the environment that will execute commands. Distinguish installed skills from effective workspace access. Ask your agent to explain a failed check before repeating work or expanding the scope.