Hire 16 AI agents.
Pay like one teammate.

Runs a read-only passive scan against any website you own — covering all OWASP Top 10 categories, headers, SSL/TLS, DNS, and cookie security — and grades it A+ to F. An optional Deep Scan runs the AI in depth mode to re-analyze findings with attack-chain reasoning and exports a formatted PDF report.

• OWASP Top 10 + headers + SSL/TLS + DNS + cookie checks
• Letter grade (A+ to F) plus an AI executive summary
• Optional Claude Code Deep Scan with live SSE stream and PDF export

The Vulnerability Scanner Agent connects to any Windows or Linux server via SSH (password or key, AES-256-GCM encrypted at rest). It runs seven groups of read-only diagnostic commands — network, processes, users, services, security events, patches, persistence — and OpenAI returns structured findings with severity, evidence, and recommended remediations.

• SSH-based — no agent or daemon installed on the target
• Seven command groups including registry persistence and event-log triage
• AI risk score (0–100), severity-tagged findings, and exec summary

Pick a topic. The coach asks 4–5 questions to gauge your level, then generates a 3–6 module curriculum (15–40 lessons) tailored to you. Switch between Lesson, Q&A, Quiz, and Practice modes. Skill level adjusts automatically based on your quiz performance — up after 90%+ scores, down after weak ones.

• Skill assessment + tailored curriculum (3–6 modules, 15–40 lessons)
• Lesson, Q&A, Quiz, and Practice session modes
• Adaptive skill level + multiple coaches per user with independent progress

The Git Coder Agent scrapes github.com/trending plus an OpenAI web search for true-velocity signals. For each candidate, it generates a structured build plan from the README and emails it to you with one-click Approve / Reject buttons. On approval, a headless Claude Code SDK subprocess scaffolds and ships the build into a workspace folder, with a budget cap and unlimited auto-retry on transient errors.

• GitHub Trending scrape + OpenAI web search, deduped and ranked
• Email approval with single-use tokens; in-app approve / reject too
• Claude Code SDK build with command blacklist and budget cap

Ask any question, optionally tag a topic, pick a depth. The agent runs multi-source web search, weighs sources, and writes a structured report (500–800 / 1,000–2,000 / 2,000–4,000 words) with every source tracked and a relevance score. Download the result as a PDF.

• Three depth modes: brief, standard, detailed
• Source tracking with URLs, titles, and relevance scores
• PDF download per report; full paginated history

The Social Media Manager lets you upload markdown describing a project, define one or more free-form personalities (Professional, Snarky, Founder voice — anything you can describe), and cycles through them on a schedule. Each post evolves — a memory window of recent posts is folded into the prompt so the storyline progresses without looping.

• Free-form user-defined personalities cycled round-robin
• Configurable text + image models (e.g. gpt-4o-mini + dall-e-3)
• Memory window prevents repeats; manual / interval / cron scheduling

Paste a YouTube URL for a 4–6 sentence summary, 6–10 takeaways, topic tags, and a type classification (tutorial / podcast / review / etc.). Or track an entire channel — the agent watches the public RSS feed on schedule and auto-summarizes new uploads (with a no-backfill first-scan policy and a per-run cap so bursty creators do not overrun your budget).

• On-demand summaries with key takeaways and topic tags
• Channel tracking on manual / interval / cron schedules
• No-backfill on first scan, per-scan cap, per-channel model override

Create any number of independent feeds, each with its own keywords, schedule, and depth. Real-time articles come from a multi-source web search. Digests are generated in three depth modes: Brief (3–5 stories), Standard (5–8 stories with a Quick Bites section), or Detailed (deep-dives plus an Analysis paragraph).

• Multiple independent feeds, each with its own schedule
• Brief / Standard / Detailed depth modes
• Paginated run history with full digest viewer per run

Knowledge Base RAG indexes everything every other agent produces — research reports, ATS candidates, interview scores, meeting transcripts, BDM results, security scans, YouTube summaries — plus OneDrive folders, SharePoint, and GitHub repos. Embeddings live in PostgreSQL via pgvector; RAG returns grounded answers with cited source chunks.

• pgvector semantic search across 11 source types
• Cited answers with link-backs to the original chunks
• Per-source-type enable/disable, configurable chunk size and embedding model

A voice console that lets you control every agent by speaking. Responses stream back as text and audio.

• 57 tools spanning all agents (config, triggers, queries, conversation)
• Continuous conversation with barge-in support
• Browser-only voice (Chrome / Edge) — no extra services to set up