Q: Important question
Does the api support conversion threads or memory, or do i need to manage conversion history myself?
AND ALSO, is there a way i can control how much someone consume my own credits? For example of i give access to poppy through whatsapp, can i limit his usage? And can he see how much he have left to use?, can i decide what type of ai model he will have?
I need to know how can i can control this thing so it wont get out of hand
athena_poppyai
Jun 15, 2026A: Hey! 👋
1 - On the API and memory. Yes! The API supports both. You pick which one fits 🙌
Option A - No memory.
Each message you send (each API request) stands alone.
Great for automations or one-shot stuff.
If you want continuity, there are ways to stitch history into the prompt yourself.
Option B - Poppy holds the memory.
There's a way to start a thread once, save its ID, and pass that ID with every message after.
Poppy remembers everything in that thread for you.
For WhatsApp-style setups, the cleanest pattern is one conversation per user.
Each person's messages stay in their own thread and stay private to them ❤️
One more thing if you're giving multiple people access. There are two common setups:
- Use ONE trained board across all users. The bot stays personal through context you pass in (name, profile, etc.)
- Train a separate board per user. Deeper personalization, more setup on your end. Most charge a premium for this!
Both work through the API. If you grab a tier with API access (Tier 5+), we'll set up a 1:1 to figure out which one fits your setup best 🙏
2 - On limiting how many credits each person can use. A few options here:
Option A - If you're going the API route (like WhatsApp), the cleanest path right now is tracking credit usage per user in your own database.
Google Sheets, Supabase, whatever you like.
From there, there are ways to gate each message against whatever cap is set.
There's also a useful trick if you want even tighter control.
Instead of letting Poppy hold the conversation history (Option B from above), you can store the history in your own database too.
That way YOU decide how much past chat actually gets sent with each new message.
And that's the biggest factor in credits per message 🙌
Option B - If you're using our native chatbots (not the API), credits currently sit at the account level. We're rolling out per-chatbot credit limits soon so you'll be able to cap how much any single chatbot can use each month 🔥
3 - On the end user seeing how many credits they have left. They're using YOUR credits, not their own, so the balance lives on your side, not theirs.
That said, once we ship per-chatbot credit limits soon, end-user visibility on what's left in their bot is something we could add too. If that's important for your setup, feel free to drop a request on our feedback page so we can prioritize it 🙌 https://feedback.getpoppy.ai
4 - On picking the AI model for them. Yes! When you set up a chatbot you pick the model it runs on. So for simpler Q&A you can use a lighter/cheaper model to stretch credits 🙌
Hope that helps!