← Career OS

Transparency

Trust is our entire fundraising mechanism — there is no paywall. So we publish how the money works. Assume ~2% of active users donate ~$4/month: at 10,000 users that's roughly $800/month. Our whole cost base is engineered to land under it.

Live AI gateway (this server)

Loading…

Every AI action passes through one gateway that picks the cheapest capable model, serves cached results when it can, meters each user, and hits a hard budget kill switch before it can ever drain the account.

Planned monthly cost at 10,000 users

LLM inference
Gateway routing, hard caching, per-user caps, shared insights.
$150 – $350
Database (Postgres)
pgvector avoids a separate vector DB.
$50 – $120
Redis (cache/queue)
Pays for itself many times over in avoided inference.
$15 – $40
Object storage + egress
R2/B2 zero-egress keeps downloads free to serve.
$5 – $25
Hosting (app + workers)
Autoscale off at low traffic.
$40 – $100
Email / notifications
Transactional email on a volume tier.
$10 – $30
Job data ingestion
Free/official sources first.
$0 – $80
Auth + monitoring
Free tiers cover early scale.
$0 – $55
Target operating point
~$350 – $650
♥ Help keep it free