Zapier Webhooks Action

Zapier Telegram AI Memory: Persistent Memory for Telegram Zaps

Zapier runs each Zap in isolation — your Telegram automation forgets every user the moment a Zap completes. retainr adds persistent memory via Webhooks by Zapier: search before the AI step, store after. No code. No database.

Add Telegram memory to Zapier in 3 steps

1

Get your retainr API key

Create a free retainr account at retainr.dev/dashboard. Copy your API key. Takes 30 seconds.

2

Add Webhooks GET to search memory

Before the AI step in your Zap, add a Webhooks by Zapier GET action. URL: https://api.retainr.dev/v1/memories/search?namespace=tg:{user_id}&q={message}&limit=5. Header: Authorization: Bearer YOUR_KEY. Map the results into your AI step.

3

Add Webhooks POST to store memory

After the AI step, add a Webhooks by Zapier POST action. URL: https://api.retainr.dev/v1/memories. Body (JSON): namespace (tg:{user_id}), content (user message + AI response). Your Zap now builds permanent user memory.

Zapier Telegram memory — frequently asked questions

How do I trigger a Zapier Zap from Telegram messages?
Zapier does not have a native Telegram trigger. Use the Telegram Bot API to configure a webhook that POSTs to a Zapier Webhooks (Catch Hook) trigger URL. When a message is sent to your bot, Telegram POSTs to Zapier and the Zap fires.
How do I scope memory per user in a Telegram Zap?
Map the sender user ID from the Telegram webhook payload to the namespace field: tg:{message.from.id}. Use the same namespace in both the search and store steps so they reference the same memory pool for each user.
Do I need a paid Zapier plan to use this?
Multi-step Zaps require Zapier Starter or above. The Telegram memory pattern requires at least 4 steps: Webhooks Catch Hook (trigger), Webhooks GET (search memory), AI action, Webhooks POST (store memory). Zapier Free only supports single-action Zaps.
How do I map the retainr search results into my AI step?
In the Webhooks GET step, retainr returns a JSON response with a results array. Map the content fields from results to your AI system prompt field. Use a Formatter by Zapier step to join multiple results into a single text string if needed.
Can I use this pattern with a Telegram channel bot?
Yes. For channel bots, use the channel ID rather than a user ID as the namespace: tg:channel:{channel_id}. This gives the channel bot shared memory across all its messages rather than per-user memory.

Give your Zapier Telegram Zap real memory

1,000 memory operations per month. Free forever. No credit card required.

Start free