Sharing AI coding sessions with teammates usually means copying and pasting chat logs into a document and losing all the context. A new open-source tool called ai-transcript takes a different approach: it converts the raw JSONL session files that Claude Code and OpenAI Codex generate into standalone, browsable HTML files.
The tool ships as two Python scripts, claude-transcript and codex-transcript, and requires Python 3.8+ with zero mandatory dependencies. You point it at your session files and get back interactive HTML documents that preserve the full conversation history, including tool calls, code diffs, and thinking blocks.
A few features stand out for team use. Built-in search works across conversation turns and automatically expands matching tool calls so you can find specific code changes. A sidebar table of contents lets you navigate between multiple rendered transcripts. There's parallel rendering for batch processing when you need to convert many sessions at once. And a --redact flag strips personally identifiable information before sharing, with a --share mode that produces files safe for distribution.
The tool supports wide (1600px) and narrow (800px) layouts, custom fonts, and the option to toggle visibility of thinking blocks, tool sections, diffs, and time gaps between turns.
Similar tools exist. Simon Willison's claude-code-transcripts and a few other viewers handle Claude Code output. This one differentiates by also supporting OpenAI Codex transcripts and including the privacy and redaction features out of the box. It's early days for the project with just two commits and a single v0.1.0 release, but the feature set is already practical for anyone who needs to document or share their AI coding sessions.