Earn AI Stocks while you code
Every approved session that clears the quality bar and enters the training corpus earns you AI Stocks. Your holdings grow on your dashboard while you work.
Contribute Claude Code, Cursor, and coding-agent sessions to train better AI. Digg runs quietly beside the tools you already use, strips secrets and PII on your machine, and every approved session you contribute earns you AI Stocks.
Every day you generate exactly the data AI labs need to train coding agents, and then it's deleted. Digg captures it consent-first, cleans it on your machine, and gives you a stake in what it builds.
Every approved session that clears the quality bar and enters the training corpus earns you AI Stocks. Your holdings grow on your dashboard while you work.
Secrets, PII, and your username in file paths are removed on your machine before anything uploads. You get a receipt of exactly what was stripped.
Choose which repos are captured, review every session, and share or delete anything, anytime.
One command to install. Nothing changes about how you code.
One command installs the digg CLI. Sign in once and capture starts on its own.
Keep using Claude Code, Cursor, or Codex. Digg only records the sessions you approve, nothing else.
Secrets, PII, and your username are stripped on-device, then a server re-scan double-checks before anything is stored.
Sessions that join the training corpus earn you AI Stocks. Track your holdings on your dashboard.
Live for Claude Code and Cursor today. Digg Light and team plans extend it to more of your workflow.
One command in your terminal. It turns the sessions you approve from Claude Code, Cursor, and Codex, prompts, tool calls, edits, tests, and outcomes, into clean training data and rewards you with AI Stocks. Nothing changes about how you code.
The other side of the network: when developers approve their sessions, AI teams license clean, privacy-filtered training and evaluation assets.
Validated task trajectories: context, tool calls, corrections, and successful outcomes.
Reproducible tasks with reference outcomes and executable verification, real reward signal, not self-report.
Resettable environments across repos, terminals, and APIs, with observable state and measurable completion.
Time-ordered actions, state changes, retries, corrections, and completion signals.
Model-ready assets, never raw activity. Every asset is explicitly approved, privacy-filtered, and provenance-tracked. Partner with Digg →
Digg strips secrets and PII on-device, shows you a receipt of what was removed, and only uploads the sessions you approve. Delete anything, anytime.
Install Digg, keep coding the way you already do, and earn AI Stocks when your approved sessions join the training corpus. Your secrets stay on your machine.