Save a link, a PDF, a YouTube video, or a bookmark export to Kompl and it reads the material on arrival, then writes it into cross-linked wiki pages instead of leaving it in a folder you never reopen. Save more on the same subject and the existing pages are rewritten, so source number fifty improves the pages it touches. Every claim carries a numbered citation back to the original text, and when two sources disagree the page records both sides with a confirm or dismiss button. A knowledge graph shows how topics connect, each link backed by the sentence that states it. An onboarding wizard stages pasted text, URLs, uploaded files, browser bookmarks, Twitter exports, Apple Notes, and Upnote exports for review before anything is compiled. Settings pick the compile model from DeepSeek or Gemini, cap daily spend, and show wiki health. Eight MCP tools let Claude Code, Cursor, Codex, and other agents search, read, and cite pages. Full backups export as one zip that restores without further model calls. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Feed SupoClip a long podcast, interview, or stream recording and it hands back a shelf of captioned, vertical clips ready to post. Paste a YouTube, Twitch, or Kick link, or upload a file, and a language model from Google, OpenAI, Anthropic, or a local Ollama setup picks the strongest moments. Each clip receives hook, engagement, value, and shareability scores, so the best candidates rise to the top of the library. Face detection keeps the speaker centered in the 9:16 frame, word-synced subtitles arrive in animated caption styles, and an AI-written hook title opens every clip. The built-in studio trims, splits, and merges segments on a timeline, with caption fonts, colors, and positions adjusted in a side panel. Optional B-roll overlays and transitions add polish. Connect your own YouTube, TikTok, and Instagram accounts to post or schedule directly from the task page, and watch views and likes feed back into future clip selection. Live progress streams during processing, a library groups clips by source video, and API keys plus an MCP server let agents drive the whole pipeline. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Retinue shows who is doing what across a fleet of AI agents, on a shared board where every task is a card with exactly one holder, a written acceptance check, and a receipt chain that cannot be rewritten. Operators publish cards, agents claim them through MCP or a token-scoped HTTP API, report progress, and hand off, and an agent that is not the holder gets a refusal when it tries to write. The Home dashboard counts what is waiting, running, and blocked, charts task flow and dispatch, and shows quota gauges for Claude, Codex, Cursor, Grok, and Kimi subscriptions. A collaboration workspace draws delegation graphs, device and model timelines, module contributions, and message conversations for any selected task. Operations pages report token usage by worker, and a data health page flags stale reports and duplicate identities. Read-only importers pull session data from Claude Code and Codex, and optional Lark, Feishu, and Telegram bridges turn chat messages into intents. The interface switches between English and Simplified Chinese. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

DispatchSEO runs an SEO program through your own AI agent: it researches keywords, queues ideas for your approval, and publishes a guide every morning. Connect Claude Code, Codex, Cursor, or the Claude app over MCP and the agent mines keywords from Search Console data and Google Autocomplete, then drops each idea into a queue showing volume, difficulty, and the reasoning behind it. Approve, reject, or drag to reorder, or switch on auto mode and skip the queue. Each morning the oldest approved idea becomes a real guide, delivered as a GitHub pull request or posted straight to a self-hosted WordPress site, checked against the live search results before it ships. Home greets you with a plain-language daily briefing, a journey tracker, publishing pace, and click and impression counters. Rankings charts daily positions with thirty-day sparklines, Search Traffic plots Search Console history, and an AI visibility page shows how often Claude cites your domain. A backlink playbook lists free and paid submissions with copy ready to paste. One deployment manages many sites, each with its own MCP token. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

LIA is a self-hosted AI assistant that works out what you agreed with each person from your own mail, calendar, files, and meeting minutes, then drafts the follow-up for you to approve, edit, or refuse. Connect Google, Apple, or Microsoft accounts, and the Tools area opens mail, calendars, tasks, and contacts for hand browsing without a model call. Relations gathers each person's recent exchanges, debrief, and open commitments on one card. Record or import a meeting and LIA writes minutes from thirty built-in templates, with decisions and owners listed. Knowledge spaces index your documents in more than fifteen formats for hybrid search, memory keeps pinned rules and preferences you can edit or delete, and routines, reminders, and a daily briefing anticipate what is due. Administrators assign a provider and model to each pipeline stage, set per-user token and cost limits, and cap daily instance spend, with Grafana dashboards for monitoring. The interface covers six languages, with voice mode and a Telegram channel as options. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

AIUsage shows where your AI coding spend actually went: it reads the session logs that Claude Code, Codex, Cursor, OpenCode, Copilot, Gemini CLI, and many other assistants leave behind, then turns them into one dashboard with tokens, estimated dollars, and activity. The Overview page totals tokens, cost, active days, and sessions for today, this week, this month, or all time, and filters by tool or device. Cost and Tokens pages chart daily spending and input, output, cache, and thinking usage, and Models ranks every model by share. The Sessions table lists duration, tool count, and cost per session, and a detail view walks through each API call and the tools it triggered. Tool Calls shows which built-in, MCP, and skill actions agents use most, Projects ranks directories by spend, and Quotas tracks subscription limits. Optional sync through GitHub, S3, R2, or MinIO merges several machines, and a public leaderboard accepts aggregate totals only. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Every team wiring its own MCP servers by hand ends up with credentials scattered across dotfiles and no list of what exists; MCP Gateway and Registry replaces that with one governed front door. Register an MCP server, an A2A agent, a reusable skill, or a custom entity type once, and the Discover page lists it with tags, ratings, and search by natural language. Coding assistants and agents connect to a single authenticated gateway backed by Keycloak, Entra ID, Okta, Auth0, or Cognito, with scopes that decide who may call which tool. Virtual servers bundle tools from several backends behind one endpoint. Connected Accounts brokers third-party logins for Slack, Atlassian, and GitHub so tokens stay off laptops. Every registration is security scanned, and an admission gate holds unsafe items for review. The settings area offers audit log filtering with JSONL and CSV export, federation with outside registries, per-group rate limits, and a searchable configuration viewer. OpenTelemetry metrics and health checks are built in. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.