Attaching an app and having the attachment work are two different things, and the gap usually surfaces in front of the whole channel. A verification step in the tool set means the wiring is checkable before deployment rather than after. Pair it with inspect-runs: building an agent conversationally is fine, but you need the execution traces to find out why it answered the way it did. Note the credit-based platform pricing — the endpoint is not billed separately during early access, but the runs are.
Runbear builds agents that live where your team works. The endpoint lets you create those agents, attach apps and contexts to them, and put them into a Slack channel — which is where they run.
- Agents — list, get, create, update, plus a personal agent
- Contexts — create, update, delete the knowledge an agent draws on
- Apps and tools — search apps, attach, connect to an agent and verify the connection, list an agent's tools, detach one
- Triggers — list, get, create an external trigger, schedule a job, update or delete
- Deployment and operations — deploy to Slack including joining channels, run on demand, and inspect recent execution traces
OAuth. The connection is scoped to your organisation and least-privilege by their description, and can be disconnected at any time. The endpoint is in early access, with the tool set still expanding.
Paste a URL, then authorize — add the endpoint to your client, then approve the OAuth consent screen