A pool is a named group of endpoints; a bot set to a pool is load-balanced across its enabled members. The Default Pool serves any bot with no specific pool or pinned endpoint.

RunPod / LiteLLM endpoints that serve your bots. After you redeploy a pod, open its endpoint and paste the new URL — no worker redeploy needed.

Limit which service tiers and organizations may serve from this pool. Leave a section empty to allow all.

Where the weekly end-of-week report is emailed, and which formats to attach. Ascend always receives a copy.

Activity across every bot in this organization (shown in your local timezone; stored in UTC).

Saved end-of-week summaries (kept 180 days).

An org-wide hold blocks chat for every bot in this organization and preserves its records from the weekly wipe.

Hourly activity shown in your local timezone (records are stored in UTC). Transcripts are kept for the current week.

Re-authentication, dual approval, and a per-bot activity log for sensitive changes.

Chat with this bot using its current config. Test runs don’t count toward usage or analytics.

Up to 20 fields this bot tries to collect. The name you enter is used wherever the field is shown — prompt instructions, lead records, and reports.

Content the bot can answer from. Paste text or add a URL. Sources are saved now; indexing & retrieval are coming.

When the bot should hand off to a person, and where that goes.

Grant paid add-ons per bot or for the whole org. An org grant applies to every bot in that org.

Where captured leads are delivered — up to two destinations.

Where this bot's model lives — which model it runs and which host serves it.

The widget authenticates with one of these keys. A raw key is shown once at creation — copy it then.

How this bot reaches your users. Toggle a channel to configure it \u2014 more channels are on the way.

An active hold blocks chat traffic per its policy. Use to pause a bot for billing, policy, or legal reasons.