Inferensys
Inferensys designs AI systems around your data, tools, and operations, from AI data retrieval and RAG to AI agents, workflow automation, enterprise AI go
Built from Inferensys's public GitHub profile, United States of America. Not on Take My Repo yet, so there are no posts or updates here.
Context Management Layer for Agentic AI. Graph-based knowledge context, SQLite index, advanced FTS5 lexical/BM25 search and cross-session live memory.
MCP server for Apollo.io, search leads, enrich contacts, manage sequences & CRM from Claude Code, Cursor, or any MCP client. 27 tools covering the full Apollo.io API.
TypeScript starter kit for production voice agents: WebRTC, telephony, realtime model APIs, STT/TTS pipelines, barge-in, handoff, evals, and post-call workflows.
Driver drowsiness and distraction detection from in-cabin video: code for Bulwark-B1, a driver monitoring (DMS) video model on V-JEPA 2.1, and Bulwark-E1, an ONNX model for CPU. Eye closure, PERCLOS, yawns, phone use, eating and drinking. Code Apache-2.0; research weights CC BY-NC 4.0. Commercially licensed models: inferensys.com/contact
Agentic SOC platform for multi-agent security ops: vulnerability detection, knowledge graphs, team based parallel analysis, and GitHub integration.
Agent skill for landing page copy that reads like a person wrote it. Works with Claude Code, Claude.ai and skills.sh.
Roast and repair your MCP setup across Claude, Codex, Cursor, Windsurf, GitHub Copilot, VS Code, Cline, Roo Code, Continue, and Zed.
Multi-agent orchestration app with planning, task graphs, sandboxed execution, live agent activity, and workspace inspection.
Bounded crawl control plane for documentation ingestion, stable diffing, and retrieval-ready record exports.
Score model outputs against explicit rubrics with an Azure-backed judge, deterministic aggregation, and diffable JSON reports.
Guarded browser task control plane with Azure-backed action planning, YAML policy evaluation, approval checkpoints, and replay timelines.
Reference patterns for in-product AI assistants with typed actions, approval gates, and append-only audit trails.
Schema-first document ingestion reference with layout-aware normalization, review queues, and provenance-preserving chunks.