Dual-model analysis

Track GPT + Claude progress, monitor divergence, and know when the meta-consensus layer needs your attention.

GPT stream

GPT-5.1 reasoning

Analyzing

Running long-context prompt with guardrails...

0% complete

Claude stream

Claude 3.7 sonnet

Analyzing

Running long-context prompt with guardrails...

0% complete

Meta consensus supervisor

Continuously compares GPT + Claude deltas and decides which follow-up questions to send back to the founder.

agreement

0.78

Follow-up needed
Clarify enterprise onboarding timeline from GTM view.
Follow-up needed
Quantify attention currently spent on analyst check-ins.
Follow-up needed
Validate data freshness for revenue cohort modelling.

Analysis timeline

Every stage is persisted through Supabase for auditing.

Context sanitized

12s ago

Prompt injection guard applied

GPT-5.1 streaming

Just now

4,100 tokens processed

Claude 3.7 streaming

Just now

3,750 tokens processed

Consensus prepping

Awaiting

Divergence threshold 22%

Meta confidence

Shows when the AI needs human validation.

Signal health74%
Divergence threshold0.22
Next action: Review consensus in /review once both models hit 100%.