OpenTag
Model-agnostic AI coworker for Slack and Teams — route to 80+ models with cost savings
Last verified:
What is OpenTag?
OpenTag is an AI coworker agent that operates directly in Slack and integrates with business tools (HubSpot, Gmail, GitHub, Stripe, Zendesk, Notion, PostHog, databases) to automate recurring tasks and maintain company knowledge. It learns patterns from team conversations and proactively suggests automations, executes tasks with scoped permissions, and organizes documentation that self-updates from actual team operations.
OpenTag pricing
Pricing model: Freemium
OpenTag pros
- Operates natively in Slack where teams already collaborate; no context-switching required
- Detects repeated requests and automatically suggests automation after third occurrence, requiring only one approval
- Intelligent permission model: scoped to requestor, not shared access; channels stay invite-only
- Sandboxed, isolated execution for each run; tears down after completion
- Self-organizing team wiki that sources documentation from actual slack conversations rather than rotting documents
- Model-agnostic with Claude support; backed by Y Combinator
OpenTag cons
- Pricing information not disclosed on landing page
- Currently limited to Slack as primary interface; business tool integrations must be explicitly connected
- Requires team adoption and clear permission setup to work effectively across organization
Frequently asked questions about OpenTag
Does OpenTag have access to all my company data?
No. OpenTag only accesses tools you explicitly connect, and permissions are scoped to the person who made each request, not shared across users.
How does OpenTag learn to automate tasks?
When OpenTag detects the same request three times in succession (e.g., Monday numbers three weeks running), it volunteers to own the task on a recurring schedule. You approve once and it runs automatically.
Where does OpenTag run?
On its own isolated machine in the cloud. Each run is sandboxed, scoped to your connected tools, and torn down when finished.