Qualify before a human replies
Ask the budget, the city or the order number, validate the answer, and route accordingly.
A WhatsApp Chatbot
You Can Build in an Afternoon.
Drag, drop and publish. Twelve trigger types start a run, fourteen action types do the work, and version snapshots mean editing a live automation never breaks the runs already in flight.
12 triggers, 14 actionsVersion-safe live editsLive API dynamic listsFull run history
TriggerNew inbound messageany connected numberAsk · quick replyWhat is your budget?Under 50L50L–1Cr1Cr+
If 1Cr+Assign senior repthen place a callElseSend listingstag: nurture
POST https://api.yourcrm.com/leads · 200 OKRun history
Runs today
1,284
+9% vs yesterday
Completion
88%
1,130 finished
Recent runs
#48213 · 1Cr+Assigned
#48212 · 50LNurture
#48211 · 1Cr+Assigned
#48210 · Under 50LNurture
#48209 · 50LNurture
Edit without fear
Every run pins to the version it started on, so publishing a change mid-conversation is safe.
Debug in minutes
Run history logs every step with the data that flowed through it.
The canvas
A node palette that covers real workflows
Send a message, media, buttons or a list. Build a dynamic list at runtime from live API data. Ask a question that validates text, number, email, phone or date. Branch on the answer, call your own HTTP endpoint behind SSRF guards, add a tag, set a field, or hand to a human.
Session variables carry data through the run, alongside read-only contact and system variables and values mapped from API responses.
- 12 trigger types including keyword, template button and API
- 14 action types from message to HTTP call to handover
- Validated questions with retry prompts
- Create with AI: paste a generated flow, validated on import
Flow · lead qualification1,284 runs today
Trigger · inbound messagelive
Ask · budget quick replylive
Branch · 1Cr+ → senior replive
Else · listings carousellive
Version snapshot v7 · runs in flight pinned12 triggers · 14 actions
How it works
From zero to live,
in four steps.
01Pick a triggerInbound keyword, template button, form submission, API call or schedule.
02Lay out the logicDrag nodes, ask questions, branch on answers, call your API where needed.
03Test the runFire it against your own number and watch the run history step by step.
04Publish safelyA version snapshot is taken, so in-flight runs continue on the old version.
What is inside,
in detail.
Dynamic lists
Build a WhatsApp list at runtime from live API data, not a hardcoded menu.
SSRF-guarded HTTP
Call your own endpoints with server-side protections in place.
Validated questions
Text, number, email, phone and date validation with retry prompts.
Session variables
Carry answers through the run and map values from API responses.
Run history
Every step logged for debugging, with the data that passed through.
Create with AI
Copy a schema-aware prompt, generate the flow in your LLM of choice, paste it back validated.
Where rivals say no
The same feature,
across four platforms.
FeatureWBIZWATIInteraktAiSensy
Up to 12 trigger typesFewerFewerFewer
Up to 14 action typesFewerFewerFewer
Dynamic list from live API———
SSRF-guarded HTTP calls———
Version snapshots for safe live edits———
Competitor entries are indicative, from public 2026 pages. See the full 89-row benchmark.
FAQ
Chatbot builder questions
Anything missing, a founder will answer directly.
No. The builder is drag and drop. The HTTP action is there if you want to call your own systems.
Runs already in flight stay on the version they started with. New runs use the new version.
Yes, through the HTTP action or the REST API, with responses mapped into session variables.
Automations are deterministic: you decide every path. The AI agent handles open-ended questions. Most teams use both.
You can paste flow JSON generated by an LLM using our schema prompt, and it is validated on import.
An inbound message, a keyword, a button tap, a new contact, a form answer or a schedule.
Yes. A quick reply or a Flow answer sends the customer down a different path.
Yes. Add a delay so a follow up goes out an hour or a day later instead of at once.
Yes. A webhook step sends data out and reads the response back into the conversation.
The run is logged with the error, and you can set a fallback so the customer is not left waiting.
Yes. Each run shows the path it took and the data at each step, which makes debugging quick.
Yes. A step can assign the chat to a team or an owner and pause the flow there.
Yes. Time based and repeating sends run on their own once you set them up.
No. Live runs finish on the version they started, and your edit applies to new runs.
Yes. A contact who opted out is skipped, and stop keywords end the flow.
Start free and try it on your own number
The free tier has no timer. Connect a number through Meta’s signup and see this working with your own conversations.