OpenCMO is an open-source growth system that unifies SEO, GEO, SERP, and community monitoring.
Built for open-source projects and developer products. See where your project is discovered, discussed, and compared — then turn those signals into reports, briefs, approvals, and actions.
Turn visibility signals into growth decisions from one open-source workspace.
OpenCMO is run daily on real projects. Two live scans show what one URL turns into.
- Auto-discovered competitors: Dify, FastGPT, OpenAI GPTs
- 50 keywords tracked (e.g.
coze 和 dify 哪个好用,coze vs chatgpt custom gpts) - 326 SERP snapshots over the monitoring window
- 100 community discussions — 61 on Hacker News, 29 on Bilibili, 10 on Dev.to
- 153 GitHub developers surfaced as potential users from related repos
- Caught: "Citability score dropped 24 points" — AI-search visibility regression flagged in the next scheduled scan
- 54 keywords, 672 SERP rank checks logged over the window
- 100 community discussions across Hacker News, Dev.to, Bilibili, V2EX
- 34 actionable insights generated end-to-end
- Caught (critical):
digigrow.uk online marketing solutionsdropped from #2 to #16 in 24 hours (May 2 → May 3) — pushed as an actionable insight without anyone polling the dashboard
Try the public demo with your own URL — one scan, full pipeline, real numbers.
- It treats growth as a system, not a checklist: SEO, GEO, SERP, community discussion, competitors, reports, and approvals live in one loop.
- It is built for open-source reality: before you have a marketing team, you still need discovery, discussion, and credibility.
- It helps you act on signals: not just dashboards, but next actions, briefs, drafts, and human-in-the-loop approvals grounded in project context.
- See where your project is visible: monitor search rankings, AI-search presence, community mentions, and crawler accessibility.
- Understand who you are competing with: map competitors, keyword overlap, and community context in the knowledge graph.
- Prioritize what to do next: identify the discussions to join, the keywords to push, and the content gaps worth closing.
- Keep execution grounded: generate reports, agent briefs, and approval-ready drafts from the same project context.
- Enter your URL on the homepage — AI scans your brand, category, keywords, and competitive context.
- Monitor SEO, GEO, SERP, and community signals on a Daily / Weekly / Monthly schedule.
- Turn raw signals into context with reports, graph exploration, and AI chat grounded in project data.
- Move to execution with drafts, approvals, and prioritized next actions.
One URL triggers a 6-stage AI pipeline that builds a complete growth picture:
| Stage | Name | What it does |
|---|---|---|
| 1/6 | Context Build | Crawl your URL. Three AI specialists (Product Analyst, SEO Strategist, Community Strategist) run a 3-round debate to extract brand name, category, keywords, and competitors. |
| 2/6 | Signal Collect | Run SEO audit, GEO visibility check, community search (Reddit, HN, Dev.to, ...), SERP keyword tracking, and GitHub potential-user discovery — all in parallel. |
| 3/6 | Signal Normalize | Clean and standardize raw data: deduplicate discussions, normalize scores, align keyword and competitor records. |
| 4/6 | Domain Review | Four AI analysts independently review the signals: SEO Analyst, GEO Analyst, Community Analyst, Competitor Analyst. |
| 5/6 | Strategy Synthesis | An AI Strategy Director synthesizes all reviews into prioritized findings and actionable recommendations. |
| 6/6 | Persist & Publish | Save results to DB, generate strategic report, surface insights on the dashboard. |
After the initial scan, schedule daily / weekly / monthly re-scans to track changes over time.
- SEO Audit: Core Web Vitals,
llms.txt, AI crawler detection, and technical site health. - GEO Visibility: monitor how your brand appears in AI-native search surfaces such as ChatGPT, Claude, Gemini, Perplexity, and You.com.
- SERP Tracking: track keyword rankings over time with crawl-based or provider-based checks.
- Community Monitoring: watch Reddit, Hacker News, Dev.to, YouTube, Bluesky, Twitter/X, plus Chinese platforms such as V2EX, Weibo, Bilibili, and XueQiu.
- Knowledge Graph: explore competitors, keywords, and community connections in one visual map.
- Reports: generate versioned strategic and weekly reports, with human readouts, agent briefs, PDF export, and email delivery.
- Potential Users: automatically discover contactable GitHub developers from your product's keywords, competitors, and related repositories. Score leads by tech-stack match and reachability, then generate personalized outreach (email, Twitter DM, GitHub Issue) through the approval queue.
- Approvals and AI Chat: keep humans in the loop while using project-aware AI agents to reason, summarize, and draft.
OpenCMO includes a report system inside each project workspace. Open the Reports tab or visit /projects/<id>/reports.
- Strategic reports: full-scan analysis with positioning, competitor context, risks, and recommendations.
- Weekly reports: 7-day monitoring summaries with trend changes, risks, wins, and next actions.
- Dual outputs: every report is stored as both a Human Readout and an Agent Brief.
- Multi-agent pipeline: human-facing reports use a 6-phase pipeline instead of a single prompt.
- Graceful fallback: if the deep pipeline fails, OpenCMO falls back to simpler generation paths so reports stay available.
OpenCMO works with OpenAI-compatible APIs, including OpenAI, DeepSeek, NVIDIA NIM, Kimi-compatible gateways, and Ollama.
git clone https://github.com/study8677/OpenCMO.git
cd OpenCMO
pip install -e ".[all]"
crawl4ai-setup
cp .env.example .env
opencmo-webThen open http://localhost:8080.
Enter your project URL on the homepage to run the first scan. If no LLM API key is configured yet, a red dot on the Settings icon will guide you to the setup panel.
Tip: you can also configure API keys from the web dashboard's Settings panel without touching
.env.
For a fuller walkthrough of provider keys and optional integrations, see the configuration guide.
Frontend development (optional)
cd frontend
npm install
npm run dev
npm run buildThe dev app runs at http://localhost:5173 and proxies API traffic to :8080.
| Capability | Platforms | Auth |
|---|---|---|
| Monitoring | SEO, GEO, SERP, Community | Optional provider keys |
| Community sources (EN) | Reddit, HN, Dev.to, Bluesky, YouTube, Twitter/X | Optional |
| Community sources (CN) | V2EX, Weibo, Bilibili, XueQiu | Free (XueQiu needs cookie) |
| Community sources (stub) | XiaoHongShu, WeChat, Douyin | Pending (MCP/Docker) |
| Publishing | Reddit, Twitter/X | Required |
| Reports | Web + Email + PDF | SMTP for email |
| LLM providers | OpenAI-compatible APIs | Required |
- AI CMO strategic scan
- SEO / GEO / SERP / community monitoring
- Versioned strategic and weekly reports
- Multi-agent deep report pipeline (6-phase)
- PDF export with branded header/footer
- 3D knowledge graph
- Approval queue and controlled publishing
- Chinese platform community monitoring (V2EX, Weibo, Bilibili, XueQiu)
- Full i18n (English, Chinese, Japanese, Korean, Spanish)
- Locale-aware AI responses (LLM follows UI language setting)
- LLM retry with exponential backoff for unreliable providers
- Simplified onboarding: enter URL on homepage, no configuration required to start
- More publishing targets
- Brand voice controls
- Deeper enterprise SEO crawls
- study8677 - Creator and maintainer
- Lling0000 - Lead contributor
- ParakhJaggi - Tavily integration (#2, #3)
- BBear0115 - Bug fixes for BYOK key isolation, base_url normalization, and reports (#9)
- See CONTRIBUTORS.md for the full contributor list
- geo-seo-claude by @zubair-trabzada
- last30days-skill by @mvanhorn
- Agent-Reach by @Panniantong — Chinese platform integration inspiration
- LINUX DO — Where enthusiasts gather