How aarnavp25-sys/brand-deal-macher works
Brand Deal Matcher collects and enriches Instagram creator data, then stores it for admin management and brand-facing creator matching. Its distinct stages include capped Apify collection, eligibility filtering and metric enrichment, keyword taxonomy classification, optional LLM labeling, and product-fit matching. The sampled code shows pipeline writes and admin proxy calls to Supabase; the demo and Edge Function implementation were not sampled, so their connection is shown as unverified.
Main entry points: schema.sql Supabase data · llm_classify.py LLM labeling · process.py Metrics enrichment · index.ts Match API
Components 17
Creator collection
- Apify runner
pipeline/apify_run.py
Processing and labeling
- Profile filtering
pipeline/process.py - Metrics enrichment
pipeline/process.py - Keyword labels
pipeline/process.py - Shared taxonomy
pipeline/taxonomy.py - LLM labeling
pipeline/llm_classify.py
Matching experience
- Brand demo
demo/server.py - Match API
supabase/functions/api/index.ts - Demo website
demo/index.html
Data and administration
- Admin dashboard
admin/server.py - Admin interface
admin/index.html - Supabase data
pipeline/schema.sql - Taxonomy export
pipeline/export_taxonomy.py
Other components
- Brand user
- Admin user
- Apify
- Amazon Bedrock
Connections 19
- Apify runner → Apify: runs actor
pipeline/apify_run.py - Apify → Apify runner: returns dataset
pipeline/apify_run.py - Apify → Profile filtering: supplies profiles
README.md - Profile filtering → Metrics enrichment: keeps profiles
pipeline/process.py - Metrics enrichment → Keyword labels: classifies creators
pipeline/process.py - Keyword labels → Shared taxonomy: uses labels
pipeline/process.py - LLM labeling → Shared taxonomy: uses labels
pipeline/llm_classify.py - LLM labeling → Amazon Bedrock: classifies with
pipeline/llm_classify.py - LLM labeling → Supabase data: writes labels
pipeline/llm_classify.py - Metrics enrichment → Supabase data: loads records
pipeline/process.py - Brand user → Demo website: describes product
README.md - Demo website → Brand demo: uses website
README.md - Brand demo → Match API: requests matches
- Match API → Supabase data: reads creators
- Admin user → Admin interface: manages data
README.md - Admin interface → Admin dashboard: uses dashboard
README.md - Admin dashboard → Supabase data: proxies requests
admin/server.py - Taxonomy export → Shared taxonomy: exports lists
pipeline/export_taxonomy.py - Taxonomy export → Match API: generates taxonomy
pipeline/export_taxonomy.py