AI Analyst for GitHub Data
Connect your GitHub repositories, get a shareable dashboard with AI-generated charts and insights in seconds.
What you can analyze from GitHub
Sreniq treats your GitHub repository CSV as a normalised tabular dataset and lets you ask questions in plain English. Typical sources our customers connect from GitHub:
- CSV exports checked into engineering repos
- release telemetry and benchmark dumps stored in /data folders
- experiment results and ML evaluation logs
- scraped or generated datasets used in research code
- pricing snapshots, fixture data, and anything you version-control as CSV
How to connect GitHub to Sreniq
Sreniq uses GitHub's standard OAuth flow — read-only scopes, no write access, and tokens stored encrypted server-side. Set up takes about a minute.
Click 'Connect GitHub'
Sreniq opens GitHub's OAuth authorization page. The connector requests the public_repo scope, which lets it read CSV files from your public repositories.
Approve the scope
Only the public_repo scope is requested — Sreniq has no access to private repos, no write access, and cannot see organisation membership beyond what the public API exposes.
Pick a repo and CSV path
Sreniq lists your accessible repositories. Choose one and either supply a path to a CSV (e.g. data/metrics.csv) or let Sreniq walk the repo tree and pick the first .csv it finds.
Ask questions in chat
Sreniq parses the file with an RFC 4180 CSV parser, then your prompts drive chart and table generation alongside an AI-written narrative.
Example questions you can ask
Once your GitHub data is in Sreniq, you chat with it. Sreniq generates the right chart, table, or narrative for each question — you do not need to pick a visualisation first.
- “Plot benchmark latency over the last 30 commits in this CSV.”
- “Compare model accuracy across experiments and call out the best run.”
- “Show pricing changes month-over-month from the snapshot file.”
- “Group test failures by suite and surface the flakiest tests.”
- “Build a release-velocity dashboard from the changelog data dump.