Upload interview transcripts, run multi-agent analysis powered by IronCrew, and generate synthetic conversation data. Built for forensics, pharma, market research, and beyond.
Three core pillars that power the platform.
Ingest VTT, SRT, and plain text transcripts up to 50MB. Automatic format detection, speaker extraction, and normalization to a canonical schema.
Run 7 analysis types through IronCrew's multi-agent runtime: summary, topics, quotes, sentiment, speakers, behavioral DNA, and interviewer scorecard.
Generate realistic interview transcripts from behavioral segment profiles. Configure persona, domain, duration, and scenario for authentic VTT output.
Domain-agnostic by design. Here are three verticals already in use.
Analyze witness interviews, suspect interrogations, and field reports. Detect behavioral patterns, extract key quotes with timestamps, and score interviewer technique.
Process HCP and patient interviews. Map prescribing behavior, detect adverse event signals, and build synthetic personas for advisory board simulations.
Extract competitive intelligence, map decision journeys, and surface unmet needs from consumer and B2B interviews at scale.
The infrastructure behind the analysis.
Multi-agent orchestration runtime in Rust. Stateless compute with SSE streaming, model routing, prompt caching, and per-run abort.
DAG-based workflow engine in Rust with 96 built-in nodes. Runs complex multi-step pipelines with parallel execution, retry logic, and conditional routing.
Local-first transcription CLI in Rust. Speaker diarization, VAD segmentation, 4 providers. Audio never leaves your machine.
Document extraction engine in Rust. Transforms PDF, DOCX, PPTX, and images into structured Markdown using local ONNX models and vision-language models.
Schedule a walkthrough with our team. We will show you the full pipeline on your own transcripts.