{"slug": "building-an-autonomous-content-pipeline-with-hermes-agent-colony", "title": "Building an Autonomous Content Pipeline with Hermes Agent + Colony", "summary": "A developer built an autonomous content pipeline that uses Hermes Agent to research trending topics, generate SEO-optimized outlines, and write publish-ready blog articles without human intervention. The pipeline integrates with Colony, a Clojure daemon that orchestrates AI workers, replacing a previous Claude CLI subprocess with Hermes Agent's native web search, terminal, and file tools. The system runs a three-stage process—research, outline, and writing—using local models like Hermes3:8b for cost-effective exploration and frontier models like Claude Opus for final output quality.", "body_md": "An **autonomous content pipeline** that uses Hermes Agent to research trending topics, generate SEO-optimized outlines, and write publish-ready blog articles — all without human intervention.\n\nThe pipeline plugs into [Colony](https://github.com/maniginam/colony), a Clojure daemon I built to orchestrate autonomous AI workers for passive income projects. Hermes Agent replaces the previous `claude -p`\n\nsubprocess as the \"brain\" that does the actual research and writing work.\n\n```\nColony Daemon (Clojure)\n  └── ROI Task Queue (SQLite)\n       └── hermes-worker.bb (Babashka)\n            └── Hermes Agent (hermes -z)\n                 ├── Stage 1: Research topics (web search tools)\n                 ├── Stage 2: Generate outline (competitor analysis)\n                 └── Stage 3: Write full article (markdown output)\n```\n\nThe daemon assigns `roi-write-article`\n\ntasks. The Hermes worker picks them up, runs a 3-stage pipeline through Hermes Agent, and reports results back via Unix domain socket IPC.\n\nRunning the pipeline for the \"ai-tools\" niche:\n\n``` bash\n$ python3 hermes-content-pipeline.py \"ai-tools\" --count 1\n\n============================================================\n  Hermes Content Pipeline\n  Niche: ai-tools | Site: aileapers.com | Articles: 1\n============================================================\n\n[1/3] Researching topics...\n  Found 3 topic ideas:\n    1. The Rise of AI-Powered Content Generation Tools [high]\n    2. AI-Powered SEO Optimization Techniques [high]\n    3. Best AI Image Generation Tools [high]\n\n--- Article 1/1 ---\n[2/3] Generating outline...\n  Outline: 4 sections, ~2000 words\n[3/3] Writing article...\n  Written: 971 words\n  Saved: output/2026-05-30-ai-content-tools.md\n\n============================================================\n  Pipeline Complete: 1 articles generated\n============================================================\n```\n\nEach stage is a separate Hermes Agent invocation with tool access. The research stage uses web search to find trending topics. The outline stage analyzes competitor content. The writing stage produces publish-ready markdown.\n\nColony's ROI system previously used `claude -p`\n\n(Claude CLI) for all AI work. Switching to Hermes Agent gave us:\n\n**Local model support** — Running Hermes3:8b via Ollama means zero API cost for research/drafting. Only final polishing needs a frontier model.\n\n**Built-in tool use** — Hermes Agent has native web search, terminal, and file tools. No need to build custom tool integrations.\n\n**Model flexibility** — Can switch between local Hermes3 and cloud models (Claude, GPT) with a flag. Use cheap models for research, expensive ones for final output.\n\n**Skill ecosystem** — Hermes ships with 90+ bundled skills. The `research`\n\nand `blogwatcher`\n\nskills complement our content pipeline perfectly.\n\nThe Babashka worker (`hermes-worker.bb`\n\n) bridges Colony's Clojure daemon with Hermes:\n\n```\n;; Invoke Hermes Agent for topic research\n(defn hermes-run [prompt & {:keys [model timeout-ms]}]\n  (let [cmd (cond-> [hermes-bin \"-z\" prompt]\n              model (into [\"-m\" model]))\n        p   (proc/process {:out :string :err :string} cmd)]\n    ;; ... timeout handling, result parsing\n    ))\n```\n\nThe worker:\n\n```\nResearch (cheap, fast)     → hermes3:8b (local/Ollama)\nOutline (moderate)         → hermes3:8b (local/Ollama)\nWriting (quality matters)  → claude-opus-4.6 (Anthropic API)\n```\n\nThis keeps costs near zero for exploration while using frontier models only when output quality matters.\n\nFor those who just want the content pipeline without Colony, there's a standalone Python version:\n\n```\n# Install Hermes Agent\ncurl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash\n\n# Pull local model\nollama pull hermes3:8b\n\n# Run the pipeline\npython3 hermes-content-pipeline.py \"home-office\" --count 3\n```\n\nIt outputs markdown files ready to publish on any blog platform.\n\n**Local models are surprisingly capable for research tasks.** Hermes3:8b handled topic research and outlining well. The quality gap only shows in long-form writing.\n\n**Hermes Agent's tool integration is smooth.** Web search and terminal tools worked out of the box — no custom MCP servers or tool definitions needed.\n\n**The -z one-shot mode is perfect for pipeline stages.** Each stage is a discrete prompt → response cycle, which maps cleanly to subprocess orchestration.\n\n**Agentic pipelines benefit from stage separation.** Rather than one mega-prompt, breaking into research → outline → write lets you use different models per stage and retry individual failures.\n\n**GitHub**: [maniginam/hermes-content-pipeline](https://github.com/maniginam/hermes-content-pipeline)\n\n`hermes-content-pipeline.py`\n\n— Standalone Python pipeline`hermes-worker.bb`\n\n— Colony daemon integration (Babashka)`output/`\n\n— Example generated articleAll running on macOS with Ollama + Hermes3:8b locally.", "url": "https://wpnews.pro/news/building-an-autonomous-content-pipeline-with-hermes-agent-colony", "canonical_source": "https://dev.to/maniginam/building-an-autonomous-content-pipeline-with-hermes-agent-colony-1pon", "published_at": "2026-05-30 18:00:13+00:00", "updated_at": "2026-05-30 18:13:31.618129+00:00", "lang": "en", "topics": ["ai-tools", "artificial-intelligence", "generative-ai", "ai-agents", "ai-products"], "entities": ["Hermes Agent", "Colony", "Babashka", "Clojure", "SQLite", "Unix"], "alternates": {"html": "https://wpnews.pro/news/building-an-autonomous-content-pipeline-with-hermes-agent-colony", "markdown": "https://wpnews.pro/news/building-an-autonomous-content-pipeline-with-hermes-agent-colony.md", "text": "https://wpnews.pro/news/building-an-autonomous-content-pipeline-with-hermes-agent-colony.txt", "jsonld": "https://wpnews.pro/news/building-an-autonomous-content-pipeline-with-hermes-agent-colony.jsonld"}}