CandiSift vs Job-Eval

CandiSift is an open-source applicant tracking system for parsing resumes, matching candidates to job descriptions, and ranking applicants through a bias-audit workflow. Recruiters can create roles, upload PDF, DOCX, or plain-text resumes in batches, and review ranked results with accept or reject decisions. The screening sequence strips personally identifiable information, applies hard filters, ranks candidates, then sends those who remain to persona agents and human review. Token-cosine ranking, technical and risk personas, and synthesis contribute to that process. Before confirming a staged batch, users can view estimated processing costs; deterministic filters run before the paid LLM stage. CandiSift connects to Anthropic Claude through Agno persona agents, with an offline stub LLM available when there is no API key. It can run in Docker, uses SQLite by default, and supports a Postgres database URL. The project describes HTTP Basic authentication, security headers, rate limiting, upload validation, request limits, prompt-injection defenses, and an append-only audit log. The API includes a bias-audit endpoint. The project has no shipped accuracy benchmark; OCR, embeddings/FAISS, and skillNER are described as swaps rather than included defaults. The software fee is $0.00 per free; LLM processing may incur per-token costs.

Job-Eval evaluates how closely a resume matches a job description. Its ModernBERT scorer returns separate 1-to-5 scores for requirement coverage, seniority fit, domain alignment and ATS-style keyword overlap. It also surfaces paragraph-level evidence derived from the scorer’s attention maps. The website accepts PDF or text resumes and job descriptions or links, with no installation required. Accounts can save and compare evaluations over time. A self-hosted version batch-scores a CSV of jobs and writes a scored CSV plus an explanation file. The maker says self-hosted runs keep data on the user’s machine. The code is MIT-licensed and the model weights are open. The free plan costs nothing, billed free forever, and includes a 156M-parameter model, CPU or GPU runs, website evaluation and self-hosted batch scoring. The website is positioned for recruiters and non-technical users; self-hosting is for engineers and job hunters. API access is not available.