Activity
Mon
Wed
Fri
Sun
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
What is this?
Less
More

Memberships

LaunchBoom Community

1.3k members ‱ Free

393 contributions to Assistable.ai
Assistable Chat Widget Not Booking
Just wanted to check and see if anyone else is having an issue where the Assistable chat widget is not booking appointments. It will pull the get_availability tool, but it will not actually book the appointment. Has anyone else seen this?
1 like ‱ 2d
@Anthony Castiglia https://lily-lunge-2d8.notion.site/Multi-Calendar-Flow-Get-Available-Slots-2101668c344a8077b9e6c52209615f34
1 like ‱ 2d
here's what it'd look like in universal prompt, but you need to run a time tool before to force it to re-check: https://www.youtube.com/watch?v=wAdhjkQTBtM
Need advice...
What do you do if your prompt needs to be longer due to pathing through complex scenarios? I have a tax client with like 6 calendars and services that the agent is having to sort and path through and book on the appropriate calendar. I've used a calendar matrix and I think I've reduced every last character I reasonably can. I'm still at 13,351 characters. My concern is that I have an upcoming client with like 20 services with a good 5-6 products each per service and will have a similar scenario. The biggest challenge with length is that I'm having straightforward tools fail on me on both SMS and voice (e.g. add a tag, calendar booking), and help.assistable is telling me it's a matter of prompt length and to reduce it to 8K characters. Any advice, direction, tips, or help would be truly, truly appreciated! ...TRULY APPRECIATED. Thanks in advance!
1 like ‱ Nov 13
Note: for chat, you may also have to update conversation history custom field & dynamically include it in prompt.
1 like ‱ 2d
@Sean McGinnis If each office operates the same (same offers, etc) you could just reference a doc (kb) to map to correct calendar and use multi-cal. doesn't lengthen prompt at all.
Tool Return Mapping
Doing some work on tool initializaiton and reference mapping to help with direct API calls and even filtering data you want sent back to the assistant - should see the mapping and request builder go into production tomorrow
Tool Return Mapping
2 likes ‱ 9d
This is amazing! So glad you're bringing this in.
Quick Assistable Voice API cheat sheet (copy/paste)
Hey Assistable community — here’s a quick, unofficial Assistable Voice (Outbound Call) API cheat sheet. Disclaimer: I don’t work for Assistable and this isn’t official or supported by me. Sharing only because people want a copy-paste example and the docs are thin/outdated in places. Use at your own risk. ✅ The Voice endpoint (outbound call) POST https://api.assistable.ai/v2/ghl/make-call Headers Authorization: Bearer asst_YOUR_V2_KEY ← v2 keys start with asst_ Content-Type: application/json Copy-paste curl curl -sS -X POST "https://api.assistable.ai/v2/ghl/make-call" \ -H "Authorization: Bearer asst_YOUR_V2_KEY" \ -H "Content-Type: application/json" \ -d '{ "assistant_id": "YOUR_ASSISTANT_ID", "location_id": "YOUR_LOCATION_ID", "contact_id": "YOUR_CONTACT_ID" }' Expected behavior If it works, you’ll get: "success": true a "call_id" you can store If it fails, you’ll still get a response, usually one of these: No outbound number on assistant Contact has no phone Phone is on public DNC list That’s the whole flow for MVP voice.
0 likes ‱ 15d
4) Current limitation (good to know) There are no public v2 endpoints yet for: - call transcript download - call cost/usage - web-voice token minting So for now: make-call starts the voice session, and Assistable/Retell handles the live call + controls in the widget.
0 likes ‱ 9d
@Deb P hi Deb!
Quick Assistable Chat API cheat sheet (copy/paste)
Hey everyone — quick, unofficial Assistable v2 Chat API cheat sheet. Disclaimer: I don’t work for Assistable and this isn’t supported/owned by me. It’s just a “here’s the curl that works” post for people who want copy-paste, because the docs are thin/outdated in places. Use at your own risk. ✅ The Chat endpoint POST https://api.assistable.ai/v2/ghl-chat-completion Headers Authorization: Bearer asst_YOUR_V2_KEY ← v2 keys start with asst_ Content-Type: application/json Copy-paste curl curl -sS -X POST "https://api.assistable.ai/v2/ghl-chat-completion" \ -H "Authorization: Bearer asst_YOUR_V2_KEY" \ -H "Content-Type: application/json" \ -d '{ "assistant_id": "YOUR_ASSISTANT_ID", "location_id": "YOUR_LOCATION_ID", "contact_id": "YOUR_CONTACT_ID", "conversation_id": "YOUR_CONVERSATION_ID", "last_message_id": "0", "channel": "Live_Chat", "messages": [ { "role": "user", "content": "Hi!" } ], "input": "Hey, what is your name?" }' What you’ll get back Look for: Assistant reply: run.response.messages[0].content Next last_message_id: run.message.messageId (use this on your next call) Costs/tokens: run.usage_cost That’s it. If you can fill in the 4 IDs + v2 key, you’re good. Happy Black Friday đŸ–€
1 like ‱ 15d
@Rico Brown đŸ’ȘđŸ’Ș
1 like ‱ 10d
Anytime you want to control the start of the conversation through any method outside of the standard tagging workflow. E.g., testing your ai agent directly to see exact responses in code. There's more fun stuff to do, but it get's pretty technical very quick.
1-10 of 393
Jakin Harper
6
437points to level up
@jakin-harper-7439
đŸ’ȘTools for Voice AI - https://ship.integraite.cloud/ ▶app-x.co/vid

Active 1d ago
Joined Jun 19, 2024
Texas, USA
Powered by