Chat Transcripts Analyzer

You can paste plain text lines like "Customer: message" with optional timestamps, or upload TXT, VTT, SRT, JSON, CSV, or HTML files up to 2 MB. JSON and CSV need speaker and message fields.

No. Parsing and analysis run locally in your browser using deterministic logic. Nothing is sent to a server, and no account or registration is needed.

The tool measures the gap between a customer message and the next agent reply when both have timestamps. Without timestamps, response times are shown as unavailable instead of estimated.

Sentiment and outcome use keyword-based rules, so they are fast and transparent but not perfect. Sarcasm and domain slang can be missed, so treat results as a guide and review the messages.

You can copy a full Markdown report or download it, along with a CSV of every message with sentiment scores and a JSON file with all metrics, for spreadsheets or dashboards.

Use the topics, keywords, and escalation signals to find recurring issues, improve macros and AI assistant answers, and coach agents on empathy and speed, for example with SupportGPT.