// ABOUT
Built by engineers, for engineers.
Your best work has always lived in private repositories — internal tools, client projects, exploratory prototypes — that no GitHub profile can surface. Deemples exists to change that, without compromising a single line of source code.
// THE PROBLEM
A developer's most valuable work is invisible.
Most of a developer's output — production systems, internal platforms, client codebases — never appears on a public GitHub profile. Recruiters and hiring managers see a curated highlight reel at best; at worst they see nothing. The engineer who shipped the payment infrastructure or rewrote the data pipeline has no portable record of that work.
Skills assessments, take-home tests, and whiteboard interviews try to fill the gap, but they measure performance under artificial pressure — not the quality and consistency of real work over time.
// THE APPROACH
Evidence derived from real commits. Zero code exposure.
Connect a GitHub or Azure DevOps account once. Deemples reads 24 months of commit metadata — file paths, diff statistics, language distributions, contribution cadence — directly inside your browser or local environment. Not a single source file is transmitted to our servers.
From that metadata we derive a verified skill profile: languages, ecosystems, contribution volume, consistency, and code-ownership patterns. The profile is yours. You decide who sees it and when, and you can revoke access in under 60 seconds.
Read-only OAuth
We request the minimum scopes needed to read commit history. We never write, never clone.
Metadata only
File paths and diff statistics — no source content, no secrets, no commit messages.
Redaction pass
High-entropy strings, API key prefixes, emails, and ticket IDs are scrubbed before anything is stored.
Revoke in 60s
Pull your OAuth grant and we detect it on the next webhook, purge the derived profile, and send a receipt.
// PRIVACY
Deterministic. Traceable. Yours.
Every metric in your profile links back to the specific commits that produced it. Nothing is inferred by a model, nothing is padded, and the computation is deterministic — run it twice on the same history and you get the same result. Your profile cannot include work you did not do, and you can audit the derivation at any time.
Profiles shared with employers are pseudonymous by default — your name, email, and repository URLs are not disclosed until you explicitly make an introduction.
// WHERE WE ARE
Johannesburg, South Africa.
Deemples is built in Johannesburg — currently in private beta. We are a small, focused team shipping iteratively. The product is free for developers, forever.