Open-Source Cold Email & Warmup Platform
Warmbly 是開源的冷郵件平台:溫暖您的收件箱、發送活動、並對每個回覆進行分流。可在我們的雲端運行或自行託管整個系統。
AI 翻譯
尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證Open source alternative to Semrush and Ahrefs
翻譯準備中
取自儲存庫 README
正在載入完整 README。
檢視完整 README ↗Keep what your SaaS promises, publishes, and delivers aligned as the product changes.
Relay, by Playbook Media LLC, is building a system for the customer-facing work that follows every software change. When a release, pricing update, or new integration affects a website, documentation, customer journey, or sales material, Relay aims to identify what changed, locate the affected claims and workflows, prepare approved work, and verify the result in the system customers actually use.
Today: Relay has an open-source SEO and reviewed GitHub execution foundation. Next: commercialize website release assurance through Relay Protect, then build fact-backed content integrity through Relay Sync and coordinated commercial releases through Relay Launch. The broader operating system is a roadmap, not a finished product.
Use Relay · Company thesis · North star · Implement in order: IRFC-0001–0032 · Full technical RFCs
Relay is an independent, MIT-licensed codebase derived from OpenSEO. We retain its research foundation, license and attribution. Some package names and agent plugins still say OpenSEO. The upstream-sync workflow remains in place for reviewed changes.
The repository currently contains:
This is a substantial foundation, not a completed Protect, Sync or Launch service. The dated release-assurance audit describes the acceptance and integration gaps identified at the time of inspection; current implementation must be checked against the source and real deployment. In particular, existing checks or a merged PR do not by themselves prove a protected site is monitored or a production problem resolved.
Three customer-facing offers share one planned Product Reality Engine:
| Product | Customer outcome | Delivery boundary |
|---|---|---|
| Relay Protect | Keep agreed production pages healthy, catch release regressions, prepare approved repairs and verify that they work. | Initial paid entry. Start with GitHub-managed Next.js sites; browser journeys are a later validated extension. |
| Relay Sync | Keep public pricing, plan limits, feature and integration claims, and documentation consistent with verified product facts. | Planned expansion. Start with accepted facts and narrowly scoped Git-backed assets. |
| Relay Launch | Turn a real product change into reviewed website, docs, announcement and sales-enablement work, with accountable owners and publication receipts. | Planned expansion. Start with fact-backed, manually approved launch packets. |
The Product Reality Engine will link source observations → scoped product facts → customer-facing claims and assets → impact findings → approved work → destination verification. Source commits, deployed behavior, published statements, and measured commercial outcomes remain separate evidence. AI can help investigate and draft but cannot substitute for authority or silently publish.
The single contributor execution plan is IRFC-0001–0032:
Each IRFC links to the authoritative technical RFC and specifies the slice and acceptance test. The queue holds development work only. Historical RFC-0001–0046 remain intact for code references and design history. Do not rebuild completed features simply to satisfy a number. Company strategy and north-star vision describe why the milestones exist.
First useful customer result: a real production regression is detected, a supported repair is explicitly approved, the customer deploys it, and Relay verifies the original failure has cleared. The next milestone is a product change causing a verified published-claim discrepancy that Relay gets corrected and independently checks.
The current hosted Relay research service has a free tier with $1 in monthly credits and a $10/month Base Plan with $10 in monthly credits, plus top-ups. DataForSEO and certain agent calls consume metered credits.
Those hosted research plans are not subscriptions for Protect, Sync or Launch. Protect is being tested as a bounded per-site paid service. Proposed Sync ($299–$999/month) and Launch ($399–$1,499/month) ranges are commercial hypotheses, not available plans or promised pricing. See company.md and the pilot specification.
Self-hosters supply their own provider credentials and cover provider and infrastructure costs. See DataForSEO API key setup.
Relay supports two self-hosting paths. Both use email and password accounts: the first visitor creates the admin account, teammates join from invitation links, and MCP clients (Claude, ChatGPT, Codex, Cursor) sign in with the same accounts.
docs/SELF_HOSTING_DOCKER.md.
docs/SELF_HOSTING_CLOUDFLARE.md.For security-conscious setups, two advanced options replace the accounts with your own network controls:
docs/SELF_HOSTING_CLOUDFLARE.md.Self-hosted feature availability depends on configured integrations, credentials and deployment bindings. See Local Postgres for the optional database backend; D1/SQLite remains supported. Review the hosting guides before exposing an instance to the public internet.
The inherited MCP server and skills expose SEO research and supported growth workflows to agents. Growth tools can read evidence and draft within their permissions; they do not grant approval, merge, or deployment authority. Read-only change-impact APIs, event subscriptions and the broader Product Reality Engine MCP surface are planned, with permissions and costs defined in the implementation RFCs.
The bundled MCP configuration points to https://relay.playbookmedia.io/mcp. Plugin branding and parts of the plugin README still use OpenSEO. The upstream MCP setup guide explains client setup; use the Relay endpoint above, or your self-hosted endpoint and its supported authentication.
Follow Local Development for Node, the pinned pnpm version, environment configuration, and local database migrations. Common commands after setup:
pnpm dev
pnpm test
pnpm types:check
pnpm format:check
Read AGENTS.md, the contribution guide, and the maintainer documentation. Start with the sequential implementation queue. The historical RFC index preserves the detailed specifications, earlier engineering decisions, and live-acceptance caveats.
Retain upstream attribution and Playbook UI seams when integrating reviewed source improvements. GitHub detachment does not disable the existing upstream-sync workflow. See LICENSE for the MIT license and upstream copyright. The shared app theme adapts Kaneo's CSS, with its MIT notice.