Evidence-Grounded AI
for High-Stakes Career Applications.
An audited engineering workspace designed to keep generated career claims anchored in verified candidate evidence. Backed by 8 live FastMCP tools, GitHub code proof, and adversarial interview defense.
Candidate Fit vs. Target JD Benchmark
Agentic AI / AI Agents
STRONG MATCH"Architected an Agentic AI workspace using Next.js, FastAPI, and LangGraph, implementing a ReAct agent capable of autonomous reasoning, multi-step tool orchestration"
Framed directly with quantifiable performance metrics to prove authentic day-1 execution capability.
Five Core Pillars. Zero Overclaims. Real Proof.
Every feature earns its place. No vanity metrics. No overclaiming.
Evidence-Backed Job Fit
Each skill from the JD matched against your resume with actual evidence. STRONG_MATCH, PARTIAL, TRANSFERABLE, or MISSING — with proof.
Deep Web Recon & Portal Intel
Deep 8-page crawl via Tavily Advanced + Jina AI Reader. Live workplace signals across Naukri, AmbitionBox, Indeed & LinkedIn with 3-way categorized evidence.
Interview Defense & Vendor Shield
Audits your vacancy with the Job Seeker Reality Check. Flags third-party staffing payrolls with verbatim quotes, detects hidden salaries, and generates adversarial interview defenses.
Live GitHub Code Proof (Tool #7)
Queries candidate repos via GitHub MCP. Grounds claims in real commit SHAs, file trees & repos for candidate trust.
Verified Hugging Face Weights (Tool #8)
Queries candidate models via Hugging Face MCP. Validates actual open-source weights, GGUF/LoRA quants, and 18,106 Total Downloads (16,559 on flagship) to audit uncorroborated AI claims.
Built Different. Architecturally.
Not a prompt wrapper. A full evidence pipeline with MCP Server, web search grounding, and claim validation.
Security & Trust Guardrails
Built for professional integrity. Not a toy.
Architected by Ambuj Kumar Tripathi
Independent GenAI Engineer · Ex-British Telecom Automation · Gorakhpur, India
“The LLM is the least reliable part of your entire stack. That is why CoverCraft was engineered with deterministic ATS keyword math, circuit breakers, and MCP tool verification instead of trusting raw generative models.”
CoverCraft was developed as an open-source, evidence-grounded alternative to superficial AI wrappers. Ambuj's architectural work on production GenAI reliability, agentic RAG workflows, and multi-stage verification has been officially featured in UptimeRobot's global engineering spotlight.
Accredited on the public M8ven MCP Trust Index with Score 89/100 (Grade B Emerging), zero security findings, and verified FastMCP stdio/SSE schemas.
Featured in UptimeRobot Spotlight. Multi-agent hybrid RAG system designed for complex Indian financial & tax circulars.
Production micro-frontends, LLM evals, agentic workflows, and live interactive AI demonstrations.
This platform charges $0.00. There are no paid tier upsells, no monthly subscription traps, and no credit card required. Developed strictly as public AI infrastructure by Ambuj Kumar Tripathi.
We make zero false claims about “ranking your CV #1 on ATS” or “gaming recruitment algorithms.” Anyone selling “guaranteed interviews” is peddling snake-oil. Hiring decisions are made by humans based on real engineering merit.
Pure, honest, evidence-grounded writing assistance. We analyze the job description, cross-reference your actual resume, research real company developments with Tavily, and prepare you for adversarial interview questions.
Explore the Engineering Behind It
Every design decision documented. Every MCP tool explained. Full source code available for inspection.
Verified Technical Ecosystem & Production Trust
Every claim is anchored in verifiable engineering proof. CoverCraft's FastMCP tool server, cloud SLA, and open-source models are audited across leading registries:
M8ven Code Audit & Trust Index
Evaluated across Glama (Grade A - A), Smithery (Verified 8 Tools), and the public M8ven MCP Trust Index (89/100, Grade B) via automated static code audit with 0 security findings and continuous commit verification.
Featured by UptimeRobot
Spotlighted on the official UptimeRobot Community Blog for high-availability cloud architecture, 3-tier GenAI circuit breakers, and sub-second deterministic fallback telemetry.
Indian Legal LLM Suite
Suite of 6 fine-tuned, GGUF-quantized models (Llama 3.2 1B · 3B · 8B) achieving 16,400+ downloads on the flagship legal build for resource-constrained, local edge inference.