FoxlyFlow
Developer API · built by Aziz Khan
The problem
Every messaging channel ships a different API.
What I built
Provider-neutral HTTP API for sending and receiving messages, contract-first with OpenAPI 3.1. First connector is WhatsApp via linked-device sessions; designed for Cloud API, SMS, RCS and email next.
Engineering decisions
- Requests are limited per IP, per API key, per workspace and per channel before any message is accepted.
- Each phone number has its own sending limits: warm-up for new numbers, a daily cap on new contacts, and quiet hours in its own time zone.
- A message that does not fit the limits stays queued, never dropped. Idempotency keys and signed webhooks with retries make every send a promise.
- One Go binary runs as both API and worker. Postgres is the source of truth and Redis holds only transient state, in a 64 MB image.
Stack
- Go
- Fiber
- OpenAPI
- PostgreSQL
- Redis
- Asynq
Written up in detail
- How to Take an AI-Built App from Prototype to Production
A working checklist for apps built with AI coding tools: secrets, data access, config, payments, background work, deploys and monitoring.
- Why AI Agents Fail in Production and How to Fix Them
Eight ways AI agents break once real users and real data arrive, how to recognize each one, and the fix that holds.
- Designing sending limits for a WhatsApp messaging API
Warm-up curves, quiet hours, new-contact caps and why a held message should queue, not drop. Notes from building FoxlyFlow.
Need something like this? See AI Deployment & Infrastructure.
More case studies
Have an AI product to ship?
Tell me what needs to be built, fixed, integrated or deployed. You work with me directly, the engineer who reads the code and ships it.