What is Zolra
Zolra is not a chatbot. It's the hybrid AI engine that works on your startup while you sleep.
The three layers of Zolra
Layer 1 — User Intelligence Model (UIM)
Zolra builds a 12-dimension model of every user — not just profile fields, but behavioural patterns, decision velocity, network engagement, and traction signals. This model is rebuilt every 24 hours and cached in Redis. Every Zolra recommendation starts here.
Layer 2 — Behavioural OS
Most platforms notify you. Zolra models your state as a founder and delivers the right signal at the right time — pressure when you're drifting, relief when you're overloaded, momentum when you're close.
Layer 3 — Hybrid scoring
Zolra uses two kinds of scoring. Deterministic scoring for profile completeness and pattern matching — fast, predictable, no LLM cost. LLM scoring (Azure OpenAI gpt-5.4 at temperature 0.1) for co-founder compatibility, investor fit, job matching, and mentor matching — nuanced, contextual, and clamped to 0–100 to resist prompt injection.
What Zolra does overnight
- —Rebuilds your 12-dimension intelligence model from the last 24 hours of activity.
- —Scans 50+ India-specific failure patterns against your current startup state.
- —Rebuilds the ecosystem knowledge graph to discover warm investor paths.
- —Drafts thesis-matched investor introductions and queues them for approval.
- —Refreshes your Morning Brief in Home.
What Zolra is not
- —Not a chatbot you have to prompt — it works proactively.
- —Not a generic LLM wrapper — scoring is grounded in your data and a curated pattern library.
- —Not a black box — every score has a breakdown, and Zolra never executes without your approval.