Win the right work, with proof your agent can cite

PIIK reads the tools you build and ship in—GitHub, deployments, incidents—and extracts what you actually did, when, and at what scale. Structured proof, not a resume claim.

Every artifact is mapped to the contribution it represents—load reduced, reliability improved, incidents resolved—so your agent can cite impact instead of adjectives.

Contributions are linked to the capabilities they demonstrate: backend architecture, performance optimization, production operations. Every capability points back to its evidence.

PIIK scores your evidence against a specific requirement and breaks the match into Evidence Fit, Semantic Fit, Seniority Fit, and Constraint Fit—each inspectable.

The same evidence is exposed to your agent as the piik_score_upwork_job and piik_get_evidence tools, scored deterministically so the same job and evidence always give the same answer.

Connect Claude, Cursor, or your own AI agent. PIIK supplies the grounded context; your agent does the drafting, proposals, and interview stories.

Connect your OpenRouter key. Model usage is billed separately through your account, and your key stays under your control.

PIIK controls what each agent may read, draft, or submit. Tokens are scoped and revocable, and you approve anything consequential before it happens.

No black box. The same job and the same evidence always produce the same score, with the four components visible and inspectable.

PIIK reads repository data only—never pushing, editing, or deleting. It never asks for Upwork OAuth, credentials, or cookies, never scrapes the marketplace, and never submits a proposal.

A single source cannot confirm authorship or production impact. PIIK blends evidence across your connected tools and tells you how confident it is, instead of overstating certainty.

PIIK surfaces evidence you connected; it does not verify authorship or production impact on its own. A gap stays a gap until you close it, and your agent is told when a claim is unsupported.