You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A community collection of real-world Hermes Agent use cases — each one backed by public evidence.
This Catalog
Official Docs
X Threads
Reddit
Organized by category
✅
Partial
❌
❌
Runnable demos included
✅
❌
❌
❌
Primary source required
✅
✅
❌
❌
Community-maintained
✅
❌
❌
❌
Note
Evidence rule. Every use case here links to a primary source: an official Hermes doc, a Nous Research companion repo, a GitHub issue with deployment details, or a first-person blog post. If the only evidence is a tweet or "community build" marketing page, it lives in research/unverified.md instead — not in the catalog below.
A scheduled agent that researches topics every morning and delivers a summary to Telegram or Discord. The canonical "hello world" of Hermes automation.
Always-on shared bot for a small team — coding, research, ops — with per-user sessions, Docker-isolated terminal, and an allowlist. The gateway also supports Discord, Slack, WhatsApp, Signal, Matrix, Email, SMS, and more; the Telegram flow is the cleanest starting point for any of them.
A community voice gateway + Dashboard plugin that keeps a continuous, interruptible voice session open while durable Hermes /v1/runs execute in the background — bounded progress, steer/stop by task ID, completion notices after reconnect.
Desktop Hermes + Telegram Topics that wires Apple Health, Threads (via open-source skill), Gmail, and Google Calendar into a single chat surface with cross-domain queries.
End-to-end autonomous novel, art, and audiobook production. Nous Research's own pipeline shipped a 19-chapter, 79,456-word book with an audiobook and a website.
Feed Hermes a raw screen recording and get a finished tutorial video on HeyGen with your AI avatar — script generation, video production, and memory of preferences for repeat work.
Install, launch, and run ComfyUI node-based media generation workflows on demand — from Stable Diffusion to video pipelines, with automatic custom-node installation and deterministic output management.
Hermes writes designed HTML compositions (via Open Design), then renders to MP4 via HyperFrames — GSAP animations, shader transitions, TTS narration, all code-driven with no timeline editor.
Hermes as one of four interchangeable executors (Claude Code, Codex, Gemini, Hermes) inside a third-party multi-agent orchestrator. Bash-bridge adapter pattern with shared identity.
Use Matt Pocock's open-source skills (/grill-me, /to-prd, /to-issues) for planning and scoping, then Hermes sub-agents for parallel implementation of vertical-sliced issues — with HITL gates to stop AI slop before it ships.
A monitoring and improvement loop that reads session history, ranks weak skills, proposes patches or new skills, invokes self-evolution, and reports learning progress.
The v0.15 /kanban slash command turns a Slack message into a durable multi-agent swarm with named workers, a verifier, and a synthesizer — persistent via ~/.hermes/kanban.db.
Hermes keeps the judgment, a deterministic runner keeps the receipts: repeatable jobs graduate from chat into checked .nika.yaml DAG files — audited before any token is spent, budget-capped, with tamper-evident run traces. Offline demo, zero keys.
Discover and call specialist agents across machines or frameworks, or expose Hermes as an authenticated A2A service with streaming, audit logs, and anti-loop controls.
Replace or augment native memory with a dedicated cognitive DB — vector + knowledge graph + temporal decay + contradiction detection. Integrates over MCP, so no Hermes plugin needed.
Wire Hermes into an Obsidian vault for persistent knowledge — read notes, create daily dashboards, cross-reference tags, maintain a self-improving personal knowledge base.
A seven-layer local memory stack with session search, trusted facts, hybrid vector retrieval, automatic extraction, a self-curating wiki, and ground-truth rules.
A read-only dashboard plugin that turns prompts, tool calls, failures, approvals, model switches, cron runs, and memory hits into auditable agent journeys.
Records a run at the process and socket boundary, replays it offline byte for byte with the network off, and forks it from any checkpoint onto a different model. Needs no plugin and no change to config.yaml.
A Kali-backed Hermes skill for authorized security assessments with scope confirmation, playbooks, zero-findings fallback, approval gates, and structured reports.
Run Hermes behind out-of-process network, filesystem, syscall, and inference policies while retaining approved memory, skills, cron, and gateway capabilities.
Autonomous weather prediction-market trading — Kelly Criterion position sizing, EV scoring against market odds, self-calibrating with real traders reporting $300→$123K in 3 months on a $5 VPS.
Self-hosted double-entry ERP and accounting run from chat. Invoicing, inventory, payroll, and multi-company books, written to a local database by deterministic skill actions.
Hermes Startup: a local-first skill that turns your personal profile into 3 of 10 evidence-labeled ideas, one bounded next action, and a prepaid per-call tool balance — no subscription, no income guarantee.
Hermes probes your local inference server, identifies the running model, detects context window, and auto-configures — eliminates first-setup friction for Ollama, llama.cpp, MLX, and vLLM.
A 10-minute copy-paste script (54★ gist) that wires Browser Use's Browser Harness as a Hermes skill on a consumer VPS — symlinked skill, OpenRouter, Browser Use Cloud free tier.
First-class provider IDs for the three major enterprise model gateways — Bedrock Converse API + Guardrails, Azure Foundry endpoint, NVIDIA NIM cloud or self-hosted. For regulated / procurement-constrained shops.
Dedicated China-facing provider IDs for Kimi/Moonshot, MiniMax, and z.ai/GLM — auto-probing endpoints, no hand-wired base URLs. Pair with Feishu/WeCom for a fully domestic stack.
The maintainers' recommended default — one OAuth login unlocks 300+ models plus a Tool Gateway (web search, image gen, TTS, browser) and Nous Chat. Scoped JWTs, no per-provider key juggling.
Reuse existing coding-tool subscriptions as Hermes model backends — Copilot direct + ACP, Codex device-code OAuth, Claude Max extra-usage OAuth — with native auth and 401 one-shot recovery.
Bootstrap Hermes with model routing, local inference, search, vector memory, automation, notifications, observability, and monitoring through a tiered Docker stack.
Run a single query across Reddit, X/Twitter, YouTube, Hacker News, and Polymarket via a single Python engine. Firefox-cookie auth for X, no API key needed on WSL. Delivered as a unified markdown brief to Telegram or local storage.
See CONTRIBUTING.md. The short version: one primary source minimum — official docs, a Nous companion repo, a GitHub issue with real deployment detail, or a first-person blog post. X posts and "community build" marketing pages don't count as primary evidence.
For Contributors
Every PR that touches usecases/*.md triggers automated verification. Learn how it works and how to pass it:
docs/verification-criteria.md — how scoring works, what's green vs yellow vs red, and how to run the script locally before submitting
python scripts/verify-usecase.py --usecase usecases/my-topic.md --check all — generate your own verification report
Please only submit use cases you have personally run (at least once) or that you can trace to a primary public source. No crypto use cases.
License
MIT for code and configs. Docs and diagrams under CC BY 4.0.
