Hi there, this looks like the timesaver I have been looking for. Looking forward to your response:

1. How is the context window charged against credit consumption? (e.g. Multiple docs, URLs, ... as permanent context)
2. Will it be possible to add "output" from a workflow straight into the context for another workspace/agent (e.g. upon approval)?
3. Can context be removed and will the agent re-train automatically in that case?
4. Do you offer custom output formats (e.g. json object(s)) as well as custom response codes for e.g. webhooks which require validation?
5. Can Agents act on rss feeds autonmously (using WebSub to receiving webhook as trigger)?
6. Please share a pointer to your API docs.

No offense here, just trying to see fit for (my) purpose.