Echoic
AI content automation SaaS
Multi-tenant SaaS that turns a brand's website into planned, rendered and published social posts.
- Role
- Full-stack developer, Symph
- When
- May 2026
- Stack
- Gemini API
- OpenAI API
- Next.js
- NestJS
- TypeScript
- Firestore
- PayMongo
- Stripe
Real interface, recorded on sample data. Names and figures shown are fictional.
The problem
Philippine brands need a steady stream of on-brand posts across Facebook, Instagram and TikTok, written in the way their customers actually talk. A single model call produces generic copy and images that drift from the brand.
How it was built
I architected Echoic's content pipeline as a chain of specialised model steps instead of one prompt:
- Crawl the brand's site, then Gemini extracts brand context and asks intake questions.
- Gemini plans the campaign and writes captions and hashtags per post.
- A slide planner lays out each carousel, and an OpenAI critic reviews the plan before anything renders.
- Gemini image models render every aspect ratio, followed by text-fidelity checks, logo compositing and brand tinting.
- Dispatchers publish to each platform, with usage and cost tracked per model call.
Around the pipeline: subscriptions, credits and paid-feature gating through PayMongo and Stripe.
Crawl
Read the brand's own site
ExtractGemini
Brand context and intake questions
PlanGemini
Campaign and post plan
WriteGemini
Captions and hashtags
Lay outGemini
Slide plan per carousel
CritiqueOpenAI
Review the plan before rendering
RenderGemini image
Every aspect ratio
Check
Text fidelity, logo, brand tint
Publish
Facebook, Instagram, TikTok and more
Where it landed
Echoic runs in production at echoic.symph.ai. The screens here are the real interface running on sample data for a fictional coffee brand.






