Skip to content

Scoring

Typelets lets you score a candidate against the problem’s rubric criteria, rather than relying on a single gut-feel verdict.

A problem defines named criteria (for example “Correctness”, “Communication”, “Code quality”). After the interview, the interviewer scores each criterion on a 1-5 scale with optional notes, plus an overall notes field. Scores are saved on the workspace and visible only to interviewer-side roles.

Scoring is entered by a human. Typelets does not auto-grade - the score is the interviewer’s judgment.

  • owner / admin / interviewer can enter and view scores.
  • candidates and viewers never see the rubric or scores - those fields are stripped before the data reaches them.

A workspace is scorable once it has a prompt and either a rubric or at least one criterion (i.e. a problem has been applied).

Through the MCP server, score_against_rubric returns the recording timeline together with the rubric and criteria, so an AI assistant can help an interviewer reason about a score. It provides input - it does not write the score. The final number is always the interviewer’s.

Some deployments can sync the resulting scorecard to an applicant tracking system. This is an optional, deployment-gated integration; if it is not enabled in your deployment, the ATS controls are not present.