Moneat
Moneat is the open-source observability platform that unifies error tracking, session replay, performance monitoring, logging, uptime checks, synthetics, product analytics, and AI observability into a single self-hosted application — replacing Sentry, Datadog, and Statuspage with one deployment. The Sentry SDK compatibility layer accepts data from @sentry/browser, @sentry/node, @sentry/react, @sentry/nextjs, sentry-sdk for Python, sentry-kotlin, sentry-java, sentry-android, sentry-cocoa, sentry-go, sentry-ruby, and Sentry.NET by updating one DSN endpoint. Datadog Agent compatibility redirects existing fleets by setting dd_url, and native OpenTelemetry OTLP ingestion accepts logs, traces, and metrics from any exporter or Collector. Error monitoring groups exceptions with smart deduplication, session replay records DOM-based user interactions linked to errors, distributed tracing visualizes transaction and span breakdowns with live service maps, and continuous profiling renders flamegraphs in pprof, JFR, and Sentry formats. Uptime monitoring runs HTTP, TCP, and ping checks with public status pages, while synthetics executes API tests, multi-step workflows, SSL checks, and DNS probes. Custom dashboards support drag-and-drop widgets with Grafana import, product analytics provides funnels and retention cohorts, release tracking surfaces crash-free rates with source map upload, and AI observability traces LLM calls end to end. Built on Kotlin and Java with ClickHouse for analytical storage, PostgreSQL for relational data, and Redis for caching, deployment uses Docker Compose with an interactive installer automating secrets and service orchestration. 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.
CodeRail
Close the laptop, walk away, and your coding agents keep working; CodeRail is the remote control that lets you supervise them from a browser or your phone without handing your machines to anyone's cloud. A small bridge on each computer connects outbound to your server, so there are no open ports and no VPN to babysit, and each machine keeps its own Claude Code and Codex logins because CodeRail never holds your AI keys. Chats show plans, tool calls, models, modes, and background tasks, and when an agent wants to run something sensitive, the approval lands with you wherever you are, including as a phone push notification. Real terminals outlive the browser tab, files can be viewed, edited, uploaded, and downloaded in place, and git work, pull requests included, sits beside the conversation that produced it. Each chat even gets a browser on the computer itself, shared live with everyone in the chat and drivable by the agent. Roles run Owner, Operator, or Viewer per computer, narrowed down to folders and agents, with every change audited, and sessions keep running with every browser closed while history stays readable offline. 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.
Exameow
Upload a PDF, a slide deck, or a stack of lecture notes, and Exameow turns the material into exam questions you can actually practice with. The generator reads PDF, DOCX, XLSX, PPTX, EPUB, ODT, TXT, CSV, HTML, and image files, then writes single-choice, multiple-choice, true/false, fill-in-the-blank, and short-answer items through any OpenAI-compatible API you point it at. Counts per type, difficulty, output language, and chapter filters shape each batch, and large documents split automatically with duplicates removed. Practice runs sequentially, shuffled, as a timed mock exam assembled from type quotas, or against your wrong-question list, which clears items after repeated correct answers. Exam mode hides answers for honest attempts; flashcard mode reveals them for review, and AI grades short answers against reference answers with manual regrading available. Teachers publish a bank to the online exam relay and share a six-digit code: students join from any browser, sit a timed session, and get instant scores on objective questions. Question banks import and export as XLSX or CSV with smart column mapping, and practice records stay in your own database. 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.
Subpool
Your company pays for AI subscriptions nobody can see, scattered across personal accounts and team plans, and Subpool is the self-hosted control plane that pools that authorized capacity and hands every employee one key with limits you govern. Administrators combine subscription and API accounts into pools behind a single managed endpoint, then allocate employee-specific keys across whichever provider accounts are healthy. The console watches account health, remaining subscription capacity, and when each plan window resets, and routing prefers subscription capacity first, falling back to paid API accounts under the very same employee key when a window runs dry. Rate-limit, expire, or revoke any person's access independently, and keep every key-to-account assignment visible and auditable instead of buried in a shared credential nobody can trace. The employee API speaks OpenAI-compatible Responses and Chat Completions, so existing tools point at your pool unchanged, while aggregate input and output token telemetry is tracked per key in provider-native units. For OpenAI accounts, a Costs API estimate shows the remaining grant balance; Bailian Token Plan accounts display their five-hour, weekly, or monthly quota windows. Upstream credentials stay encrypted, and prompts, responses, and source code are never persisted. One Go service with an embedded console and PostgreSQL runs from Docker Compose, with sign-in on port 8081 and the employee API on 8080. 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.
DiceFrame
Game night keeps dying in the group chat because nobody can be in the same room, and the one willing game master is burned out by October. DiceFrame puts an AI game master on your own server instead: open the web UI, describe what your character does in plain language, and the engine adjudicates the action, rolls the dice check once, applies the result to your character sheet, and narrates what happens next. Run a solo campaign at midnight or invite friends through a link and play the same shared world from different cities, with waiting states, away players and live updates keeping the table in sync. Rules are layered rather than decorative; D&D 5e-inspired d20 play with advantage and disadvantage, CoC 7e-inspired d100 with bonus and penalty dice, custom d20 systems and freeform no-dice narrative all track hit points, gold, items, experience, death and revival properly. Build worlds, lorebooks and adventures in the content library, import scripts, PDFs and Foundry or Fantasy Grounds modules, or let AI assistance draft the world, rules and characters for you. Character cards are reusable across games, long campaigns are held together by session summaries and optional memory recall, and the story can be read aloud with text-to-speech through system voices or your own endpoints. Plugins and chat bots extend the table, all through your chosen OpenAI-compatible model endpoint. 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.
KaletaCMS
Kaleta is a business-website CMS where the page you edit is the page your visitors get, not a preview of it. The visual builder works on the live canvas: add sections, drag elements, edit text in place, and style every element separately for desktop, tablet, phone, and hover before you publish. Forty-six elements and forty ready-made sections cover pricing tables, timelines, before-and-after sliders, galleries, accordions, carousels, maps, and mega menus. Consistency comes from a design-token system: change one colour, font, spacing, or radius token and the whole site follows, with a WCAG contrast check on every colour pair and export in the W3C token format. Blocks you reuse become components edited once and updated everywhere, and each page carries its own search title, sharing image, and visibility, with the last twenty publications kept as versions. Moving in is a first-class feature: importers take WordPress exports including WooCommerce products and Polylang or WPML translations, plus Ghost, Blogger, Joomla, Drupal, and Webflow, creating a redirect from every old address so rankings survive. An optional AI assistant connected through MCP builds pages and handles chores only as drafts you approve. Forms, booking, enquiries, media folders with focal-point cropping, multilingual publishing, and GDPR duties such as automatic personal-data deletion are built in rather than bolted on. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. GPL-2.0 licensed.
BlogFactory
BlogFactory is the first open-source agent control plane purpose-built for multi-site content operations, shipping 16 distinct capabilities across four operational pillars — Operate, Grow, Control, and Deliver — in its v0.1.0 release. The platform keeps source evidence, AI-generated drafts, revisions, SEO metadata, human review, preflight checks, and CMS delivery destinations in one workspace while enforcing a server-side authority ceiling that prevents agents from ever publishing live content, accessing credentials, or performing administrative actions. AI agents connect through a Streamable HTTP MCP endpoint with OAuth authentication and site-scoped tool discovery, compatible with Codex, ChatGPT, and any MCP-capable client. Content creation supports single articles from keywords, titles, URLs, PDFs, text, or YouTube sources, batch campaigns, and programmatic content via reusable templates with structured data rows. The Search Growth module connects Google Search Console complete-day data to ranked optimization opportunities, 30-day growth plans, content classification by signal type (growing, low-CTR, almost-ranking, page-two, zero-click), indexing checks, and semantic internal-link suggestions. Brand Voice separates global voice rules, writer profiles, knowledge context, and brand guidelines from article mechanics. Approved drafts are delivered as CMS drafts to WordPress, Ghost, Wix, or Framer through idempotent, optimistic-locking-protected handoffs with explicit destination selection and preflight validation. The stack runs on React/Vite, a Hono/TypeScript API, PostgreSQL, and MinIO via Docker Compose, with providers including OpenRouter, OpenAI, and Google services under a bring-your-own-AI model. 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.
PharosRAG
Your team's documents sit in folders nobody can search, and the answer to a simple question hides on page 40 of a PDF someone saved last year. Pharos ingests PDFs, scans, Word documents, presentations, and spreadsheets into a local knowledge base, then answers questions against that library with citations traced back to the exact page. Hybrid retrieval finds the passages that matter, and every answer carries its evidence so readers can verify the claim in the source. Access control decides visibility per person: an API key resolves to an identity with a tenant and group memberships, retrieval filters hard to only the documents that identity may see, and a document with no permissions attached stays invisible to everyone. Two doors lead into the same library. The HTTP API runs a closed question-answering pipeline that retrieves, grounds, and answers in one call, and coding agents connect through MCP to decide for themselves when to retrieve, how to rewrite a query, and when a question needs multiple hops. A resident daemon keeps the search index and models warm, so every session shares one backend instead of reloading heavy resources. The deployment scales from one machine to replicated application tiers behind a load balancer. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.
Deadair
Somewhere in your house, a presenter is introducing the next record, and that presenter is a language model running on your own server. Deadair builds you a real radio station rather than another playlist: it picks records from your Navidrome, Jellyfin, Emby, or Plex library, or from Spotify, YouTube Music, Pandora, and Apple Music, writes what the host says between songs, speaks it in that host's own voice, and broadcasts the result as one shared live stream over Icecast, with no skip button and the same show for every listener. From the console you steer it in plain sentences: tell it "rap like Snoop Dogg" and the next hours programme themselves to that brief, or ask why the record on air was chosen and get an answer. The forward lineup runs several hours deep, visible and editable, and nothing reaches air until its audio is already on the machine, so the stream never stutters waiting on a download. Words and voice are separate choices, swapped live in Settings: break-writing can run on a hosted model while a local Kokoro or Chatterbox voice reads it, and a station with no model at all still talks. Twenty-seven bundled plugins cover music sources, metadata, news, weather, and speech, and third-party plugins install from the console. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.
GymCoach
Strength training, cardio, and body metrics land in one workout log on your own server, backed by an optional AI coach that explains its advice from your real numbers. Quick entry such as 100x8@9 logs weight, reps, and effort in one line, with a rest timer, plate-loading calculator, and warm-up ramp beside the set table. A weight picker offers only loads your equipment can build, and a readiness check-in adjusts the suggested load and says why. Templates for 5/3/1 BBB, GZCLP, nSuns, PPL, and Upper/Lower run as written and edit like any custom program. Progress pages chart estimated 1RM, weekly volume per muscle group against MEV and MRV bands, a muscle heat map, bodyweight goals, and a training calendar. Garmin FIT, TCX, and GPX imports need no cloud account and add heart-rate charts and aerobic decoupling to each run or ride. The coach writes a weekly debrief, chats in a streaming conversation, and drafts a full program from a one-sentence goal, using your own Anthropic or OpenRouter key. History imports from Strong or Hevy CSV files with a dry-run preview and exports to CSV or TCX, and an MCP connector lets ChatGPT or other agents analyse it. The app installs as an offline-capable PWA, speaks English, French, Russian, and German, and keeps every user's data separate. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.
Mistgate
Running a VPN for family, friends, or a small team usually means babysitting servers one SSH session at a time, and Mistgate replaces that with one panel that keeps the whole fleet configured. Add a server from the browser by typing its address and root password: the panel pins the host key, checks the host, opens the ports, and installs the signed node agent for you. Each person gets one subscription link that their app turns into the server list, or a personal AmneziaVPN key per device, so a lost phone can be revoked without touching anyone else. Groups decide which protocol profiles a person receives, while per-user traffic quotas, expiry dates, device limits, and per-device speed limits keep sharing fair. Hysteria2 brings QUIC-based speed with port hopping, obfuscation, and certificate options; VLESS REALITY and AmneziaWG 2.0 and 3.1 cover networks where one protocol gets blocked, and Cloudflare WARP can serve as the exit for all three. Every five minutes the panel connects to each server the way a real client app would, a doctor on every node runs host checks with safe fixes, and alerts reach you on Telegram before your users notice. The public address shows strangers a decoy site while the admin hides behind a secret path, and encrypted backups go to your Cloudflare R2 bucket on a schedule. 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.
JellyGlance
Running a Jellyfin or Emby server usually means a dozen open tabs; JellyGlance folds your whole media server into one screen you can leave open on a wall. Live sessions show who is watching what, from where, on which client, with direct play or transcode state, codec, bitrate, subtitles, and progress for every stream. Playback history accumulates in its own database, so statistics survive even when the media server forgets: top movies, series, libraries, clients, and users, watch-time trends, activity heatmaps, and weekly, monthly, and yearly recaps rendered as shareable cards. The home dashboard arranges by drag and drop, with presets, density controls, themes, kiosk mode for a wall display, and alert rules. Recently added shelves group by library, while library views surface totals and metadata gaps. Jellyseerr and Overseerr requests appear with poster cards, requester context, and admin actions, and the release calendar pulls upcoming titles from your *arr apps beside the queues of your torrent and Usenet clients, stalled items included. Wizarr invites, Tdarr transcodes, and Maintainerr cleanup connect from the same integrations area. Notifications travel by webhook, Discord, Gotify, and email newsletters, backups export to a plain folder, and sign-in spans local accounts, Jellyfin Quick Connect, and OIDC. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.
Overclick
OverClick is the task board where your AI agents clock in: you write the assignments, they claim the work, and you stamp what actually got done. A card here is a contract, not a ticket: what should happen, why it matters, and a plain-language test script that says how to confirm it, all written before anyone starts. You tell your agent to grab the next task, and it claims the card over MCP, declares the CLI and model it really runs on, receives a self-contained briefing with the mission context and branch convention, and the card slides into progress. When it finishes, the handoff lands with a summary, evidence, branch and pull request links, and measured telemetry, the tokens per model and minutes the run actually took, read from the session transcript rather than guessed. Every card keeps three roles straight, who asked, who executed, and who reviews, and only a human can mark it validated; reopening with a comment puts your note in front of the agent's next claim. Big decisions can start as RFC cards whose approval spawns the execution cards, and insights roll up what each project, mission, and model consumed. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.
Narralume
NarraLume is a writing studio for long novels where the AI proposes and you dispose, so a book stays yours across tens of thousands of words. Everything lives around the story bible: your stated intent, outline, entities, confirmed facts, relationships, chronology, and foreshadowing, kept as the single source of truth that planning, drafting, and review all read from. You can write chapters by hand in the studio, save versions and comments, ask for a selection edit, or delegate a chapter, a run of chapters, or a whole quick-creation plan to a model; whatever comes back arrives as a candidate with diffs, never as silent edits to your manuscript or canon. Review findings come with evidence you can accept, reject, or mark as false positives, and when generated prose collides with a locked fact the task stops and waits for your ruling instead of improvising around it. The run center keeps every AI job's steps, model calls, and checkpoints recoverable, the long-form lab searches story memory and previews plot impact without touching the book, and delivery ends with quality checks, exports, snapshots, and full-library backups. 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.
Workbench
For teams running BullMQ, Workbench makes the inside of your job system visible, searchable, and fixable from a single browser dashboard. Live counters track completed, failed, active, and waiting jobs per queue alongside p50 and p95 latency and throughput, so a backing-up queue announces itself before your users do. Every run sits in a filterable table where you can expand the full payload and attempt history, retry with a keystroke, or click a line in a failed job's stack trace and land on that exact file in your editor. Failures are grouped by error class and ranked by frequency with a 24-hour trend, which makes a fresh regression obvious the moment it spikes. Parent and child jobs render as a real graph you can drill into and replay subtree by subtree, schedulers can be paused or edited without a redeploy, and new jobs can be enqueued from a form that validates the payload first. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.
AgenTerm
AgenTerm lets your AI agent work inside a remote computer's terminal with the owner's permission, through a relay server you host yourself. Your agent connects to the server over MCP and creates an invitation, single-use and expiring in fifteen minutes by default, or reusable across several devices; you privately share the generated connection command with the person who needs help. That person runs a small client on macOS, Linux, or Windows, reads what access means, and authorizes it; the client then dials out to your server, so the target computer needs no SSH service and no open inbound port. The agent can then run one-off commands with a chosen working directory, environment overrides, and timeouts, receiving stdout, stderr, and exit status, or hold an interactive session open, reading output as it arrives and typing answers to prompts. The person at the keyboard watches every command and its output in the client and can disconnect at any moment, and the client reconnects by itself after a network drop. Use it to troubleshoot a colleague's machine without swapping logs, help family through a terminal problem, or drive a development box with no agent or model account of its own. The relay itself is a single Go binary guarded by a management key, deployed with Docker Compose or systemd, with in-memory sessions and a health endpoint. 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.
AGenUI Studio
AGenUI Studio turns a written description into a working interface card: an AI agent generates the UI, binds it to your APIs and data, and packages it so it runs anywhere, no Studio required. Bring your own model by entering any OpenAI-compatible or Anthropic Messages endpoint and key in the console settings, and generation proceeds contract-first, steered by design rules you publish as versioned Markdown revisions. You preview each card live as it is generated, then refine it over multiple turns while every component, binding, and action you did not target stays protected from accidental edits. The console also manages the supporting cast: a local knowledge base, API definitions, and reusable operators, all starting from bundled demo data you can swap for real services whenever you are ready. When a card is finished, download it as a self-contained Runtime Package carrying its data sources, bindings, actions, and operators, and execute it with the bundled Card Runtime inside your own application, or register a delivery target and publish straight from the workbench over a callback or a message queue. Applications subscribe to native run events over HTTP and SSE, and production deployments can replace the local defaults piece by piece, MySQL for storage, Redis for hot state, MCP services for knowledge and data, while the packages keep the same contracts. 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.
OpenMail
Your entire mail system can live in one self-hosted suite with OpenMail, from SMTP, IMAP, and POP3 delivery to three-pane webmail and the admin console that runs it. Standard clients such as Thunderbird, Apple Mail, and phone apps connect over SMTP ports 25, 587, and 465, IMAP ports 143 and 993, and POP3 ports 110 and 995, while people use the three-pane webmail with folder tree, conversation list, and reading pane at port 3000. You read, write, reply, forward, batch-organize, drag messages between folders, filter stars, search sender, subject, and body text, and follow conversations threaded automatically by their References headers. The rich-text composer autocompletes recipients, accepts dragged attachments and pasted screenshots, saves drafts every 12 seconds, sends on a schedule, and inserts reusable templates imported from HTML or generated with AI. Any OpenAI-compatible endpoint can draft replies, translate mail into twelve languages, and summarize messages, with provider keys encrypted on your server. Administrators get real-time mail statistics with 14-day send and receive trends, user quotas and bans, multi-domain aliases and catch-all addresses, outgoing queue visibility, and audit logs. Adding a domain generates DKIM keys and copy-ready SPF, DKIM, and DMARC DNS records with one-click checks, while rule-based spam scoring, TOTP two-factor authentication, and role-based permissions protect accounts in ten interface languages including right-to-left Arabic and Urdu. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.