Character-based chat and storywriting with large language models: TavernAI is the open-source frontend that leaves model choice to you. It generates no text itself; it connects to the backend of your choice - OpenAI (including GPT-4), Anthropic Claude, KoboldAI and KoboldCpp, Oobabooga's Text Generation Web UI, NovelAI, Ollama, and the crowdsourced Horde - so cost, model quality, and content policy are decided by your backend, not the interface. Characters are defined by portable card files in PNG or JSON format with personality, scenario, and example dialogue, and tens of thousands of community-made cards from sites like Chub.ai import directly. Conversations support group chats with multiple characters, a story mode for long-form writing, message swiping to branch between alternative responses, and full editing of any message. World Info injects lore into context when keywords trigger, keeping long roleplays consistent. Themes, custom backgrounds, and configurable generation settings round out the interface. It runs on Node.js, and the SillyTavern project began as a fork of it.

Keeping users informed about software releases while gathering community feedback on upcoming roadmap items becomes effortless with ShipShipShip, a self-hosted changelog, release notes, and feature voting platform. Product managers can draft rich release announcements using a formatting editor that supports Markdown, embedded media uploads, color-coded category tags, and exact publication timestamps. Instead of juggling disconnected feedback spreadsheets, teams organize backlog proposals, active development tasks, and shipped updates across a drag-and-drop Kanban board or a filterable list view. Visitors browsing the public roadmap can upvote proposed features to help prioritize the next sprint and react to announcements using eight expressive emoji reactions. When an item moves into a shipped status, the built-in newsletter engine automatically dispatches formatted release emails over SMTP to subscribers using customizable per-event templates and one-click unsubscribe links. Design-conscious organizations can install manifest-driven public themes, map internal workflow statuses to public roadmap columns, and tailor brand colors and logos. International audiences can navigate the interface in seven languages, including full right-to-left layout support, while engineering teams automate event creation through the REST API. 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.

A training institute can run its courses, batches, trainers, and students from a single self-hosted TutorLoom installation it fully controls. One deployment serves many organizations at once, and every query is scoped to the organization, so trainers and learners only ever see their own data. Institute admins manage people and time-boxed batches, inviting trainers and students one by one or in bulk from CSV and Excel files. Courses are built as levels, chapters, and steps mixing videos, documents, tasks, and quizzes, with prerequisites and sequential unlocking so learners follow the path you designed instead of wandering. Students work through a guided step-by-step roadmap on the web or in the mobile app, watching completion fill in as they go and earning points and levels along the way. When learners submit work, trainers review everything in a single grading queue, approving or rejecting with a score and written feedback, then export progress to Excel when management asks for a report. Finishing a course produces a PDF completion certificate automatically. One account can hold several roles at once, so a trainer who also administers an institute switches portals without signing out. Branded email notifications keep learners informed under your own institute's name, and a documented REST API with interactive documentation powers the web app, the mobile app, and your own integrations. 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.

Flowplan puts your team's documents, databases, whiteboards, handwriting notebooks, and daily journal into one self-hosted page tree, so the work lives in a single place you control instead of five subscriptions. Write in a block editor that stays out of the way, and when a number appears mid-sentence, type the calculation and press Tab to drop in the result, or graph functions and vectors in 2D and 3D without leaving the page. The same records can be viewed as a table, board, calendar, timeline, gallery, list, feed, chart, form, or sprint board, with relations, rollups, and formulas tying them together, and projects get ticket numbers, sprints with burndown, and automations on top. Brainstorm on whiteboards with sticky notes, mind maps, and voting; keep a journal page per day with templates and an optional PIN lock; and write by hand in notebooks with lined, grid, or Cornell paper, annotate PDFs, and sketch inside documents with a pressure-sensitive pen. Teammates edit together with live cursors, comment in inline threads, and share single pages with guests through password-protected, expiring links, while permissions reach down to a single record. AI is off until an admin enables it, per feature and per person, and it can run entirely on your own hardware or answer questions about the workspace with links to its sources. 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.

Atlas delivers a complete business operations platform in a single Docker Compose deployment, replacing the need for separate CRM, HRM, project management, invoicing, and file storage subscriptions. The CRM module provides pipeline management with Kanban views, contacts, companies, deals, leads, forecasting, saved views, and web-to-lead forms with Gmail integration for per-user email tracking inside deal timelines. HRM handles employees, departments, org charts, leave management, and attendance tracking. The invoicing system generates professional invoices with recurring billing, customizable templates, and payment tracking. Agreements provides PDF contract generation with legally-binding e-signatures, sequential signing workflows, counterparty linking, audit trails, and automated reminders. Projects integrates time tracking, billing, client management, budgets, and reporting in one view. Drive offers file storage with versioning, sharing, comments, activity logs, and password-protected links with Google Drive import and export. The calendar syncs bidirectionally with Google Calendar across month, week, day, year, and agenda views. Write provides rich text documents with cover images, comments, and templates, while Draw adds an Excalidraw-based whiteboard canvas with PDF export and presentation mode. The platform supports five languages, multi-tenant isolation, role-based permissions, and includes an Odoo importer that ingests res.partner, crm.lead, and CRM activity CSV exports for migration. Built on React, TypeScript, Vite, TanStack Query, and Zustand on the frontend with Express, Drizzle ORM, PostgreSQL, Redis, and BullMQ on the backend. 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.

BrightBean Chat is a self-hosted chat automation platform that builds visual flows once and runs them across Instagram, WhatsApp, Messenger, Telegram, SMS, and email, so your audience automations live on your server instead of a per-contact subscription. You drag conditions, delays, randomizers, data collection, and message nodes onto a canvas, test the draft while the live version keeps running, then set it live with every save versioned. Flows start from keywords, comment-to-DM triggers, story mentions and replies, follows, ref URLs, QR codes, or calls to the public API. Every channel lands in one shared inbox where teammates assign conversations, label them, schedule replies, and take over from the bot, which pauses the automation for that contact. Contacts carry custom fields, tags, and segments that stay unified across channels, and imports and exports keep the list portable. Broadcasts and multi-step drip sequences check opt-outs and each platform's 24-hour messaging window before sending, pace themselves per channel, and show exactly who will receive a message and who is skipped. Per-node counters track sent, delivered, failed, and clicked, while a bearer-authenticated REST API and signed webhooks hand leads to Make, Zapier, n8n, or your own systems. 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.

The social preview cards shown when links hit Twitter, LinkedIn, Slack, or Discord, generated on the fly: Oxigen is a small Go service for dynamic Open Graph images, no designer required per page. Instead of exporting static images per page, you compose a card once and drive it with parameters: title, author, website, logo, background image, and background dimming all arrive as URL query arguments, so your og:image meta tag simply points at the Oxigen endpoint with the page's own values substituted in. Every blog post, product page, and doc gets a branded, correctly sized preview generated on the fly. A built-in web UI covers interactive composition - tweak the text and imagery, watch the preview update, copy the resulting URL - while the same rendering path serves programmatic API use from static site generators, CMS templates, or build pipelines. Rendering is pure Go using the gg 2D graphics library, freetype fonts, and the imaging package, built on the kyoto framework by that project's author. Deployment is one stateless container on port 80 with no database and no external dependencies, so instances scale and restart freely.

Zeabur is a self-hosted web platform that deploys and runs your services on your own Kubernetes cluster, presenting the cluster as simple projects and services instead of raw manifests. Every change starts as a readable plan listing the exact fields it will write and who owns them today, and nothing is applied until you confirm it. When something breaks, Zeabur says so in plain words and points at the fix: a worker that restarted fourteen times gets a button straight to the logs from the run before the restart. You deploy from Git repositories, from an uploaded folder or ZIP archive, or directly in the dashboard, then follow live logs, open a terminal inside a container, and watch CPU and memory move over time. Projects are namespaces and all state lives in Kubernetes objects, so there is no Zeabur database to back up, and uninstalling the platform leaves your Deployments, Services, and Ingresses running untouched. Companion pieces add passkey sign-in for kubectl with tokens that expire in minutes, image builds from Git through a BuildKit operator, and a catalog of cluster components and app templates. 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.

ClientSt0r is a self-hosted operations platform for managed service providers that keeps client documentation, IT assets, credentials, and the service desk in one multi-tenant system running on your own infrastructure. Open any ticket and the client's active contract, the asset it concerns, the runbook that covers it, and the vault entries you are allowed to use appear alongside the conversation, so nobody switches between four products during an outage. Readiness scoring grades every client against what should be documented, explains each gap, and queues the fixes, while versioned documents with owners and review dates keep runbooks from quietly going stale. The credential vault encrypts secrets with AES-256-GCM, isolates them per client, enforces IP, geography, and time-of-day access rules, logs every reveal, generates TOTP codes, checks passwords against known breaches, and imports existing Bitwarden or Vaultwarden libraries. Asset records cover hardware inventory, IP address management, rack layouts, and printable QR labels. The service desk adds queues, SLA timers, dispatch views, time entries, quotes with electronic signatures, invoices, contracts, approvals, a workflow rule builder, a threading email engine, and guided client portal forms, while procurement, profitability reporting, CVE scanning, SSL expiry tracking, and uptime monitoring round out daily operations. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Built on Elixir's fault-tolerant OTP runtime, Tymeslot delivers the complete scheduling toolkit that Calendly charges for and Cal.com relicensed away from open source in 2026. Share one booking link and let clients pick available slots — your calendar stays synchronized and double-bookings disappear automatically through real-time availability checking against Google Calendar, Microsoft Outlook, and any CalDAV server including Nextcloud. Meeting types support configurable durations, buffer times, and availability windows that clients can book without email back-and-forth. Phoenix LiveView powers the real-time UI with WebSocket connections that update instantly as bookings arrive, without client-side JavaScript framework overhead. The platform attaches Zoom video rooms to confirmed bookings through your own Marketplace app credentials, sends confirmation and reminder emails via SMTP, Postmark, SendGrid, Mailgun, or AhaSend, and supports custom booking page themes with embed options for integration into existing websites. Circuit breakers, connection pooling, and graceful degradation maintain availability during provider outages. Authentication supports local accounts, GitHub OAuth, and reCAPTCHA protection on public booking forms. Self-hosting deploys as a single container with PostgreSQL embedded — no separate database container required. Two named volumes persist uploads and database state, while the same image powers the managed cloud at tymeslot.app so self-hosted users get identical features without a crippled community edition. Deploy via Docker, Cloudron with automatic updates, or Railway from a community template. 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.

Somewhere in your home sits a box nobody has opened since the move, and Attic is the self-hosted home inventory that turns its contents from a mystery into a searchable record. You catalog appliances, tools, electronics, books, games, and furniture as assets, each with photos, receipts, manuals, purchase date and price, condition, and warranty dates attached. Nested locations mirror the house itself, rooms containing cupboards containing boxes, with photo galleries so you recognize a storage space on sight. Categories carry their own custom fields, inherited by subcategories, so electronics can demand a serial number while books ask for an ISBN. Items stored inside other items, like tools in a toolbox, are tracked as asset contents and move with their container. Collections group a game library or a moving box across locations, while tags label anything that needs repair or is out on loan. Adding stock is fast: import plugins pull metadata and cover images from Google Books, TMDB, BoardGameGeek, IGDB, and Rebrickable for LEGO sets. Search combines full-text with nested match-all and match-any filters, and saved searches pin to the sidebar as shortcuts. A history timeline records repairs and maintenance per asset, and the dashboard totals purchase value across currencies while flagging expiring warranties. OIDC single sign-on, scoped API keys, a REST API, and an MCP server connect Attic to scripts and agents. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Your front desk phone rings at midnight and nobody sane is awake; QuickVoice answers it with an AI agent you configured, on servers you own, with call recordings you control. Build agents in the web console: write the prompt, pick speech-to-text, language model, and voice, set the first message, then link a real phone number from Twilio or Telnyx through LiveKit SIP trunks. Inbound callers get a receptionist that answers questions from your uploaded knowledge base, PDFs and web pages chunked and embedded for live retrieval, while outbound mode fires single quick calls or batch campaigns for reminders and follow-ups. Agents can invoke allowlisted MCP tools mid-call, so checking an order status or booking a slot happens inside the conversation instead of after it. Every call lands in the log with transcript, recording, and metadata, searchable from the analytics dashboard, and privacy controls redact personal data by default, can suppress recordings entirely, and sweep old transcripts on a retention schedule. Multi-language voices, per-agent runtime overrides for campaigns, organization accounts with API keys and audit logs, and Stripe billing hooks round out a platform built for teams who would rather inspect their voice stack than rent a black box. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Written in optimized C++ where competitors use Python or Java, NZBGet saturates fast connections while idling on CPU and RAM - which is why the performance-obsessed Usenet downloader has long been the client of choice for NAS boxes, Raspberry Pis, and routers as much as full servers. The engineering shows in the recovery pipeline. Instead of shelling out to par2cmdline, NZBGet integrates the par2 source directly and exploits its knowledge of exactly which articles failed - quick par-verification checks only what needs checking, and its own multicore repair implementation runs 2-3x faster than the external tool on the same hardware. Fast par-rename deobfuscates scrambled filenames in seconds without a full verification pass, and rar-rename recovers names for multivolume archives even when no par2 files exist. DirectWrite assembles articles straight into sparse destination files, skipping temporary-file churn entirely, with an article cache and queue-pausing options to eliminate disk contention. Automation is complete: a remote web interface, full JSON-RPC API, RSS feeds with duplicate detection, scheduling and prioritization, and an extension manager for Python and Bash scripts triggered by download events. Sonarr, Radarr, and every major indexer integrate natively.

Pretty Please Print turns one office's shared 3D printer into an invite-only request board: colleagues upload a model and follow it from request to delivery. Requesters upload an STL or 3MF file, or start from a Printables link the owner enabled, and the mesh renders in the browser, auto-framed and draggable, so everyone checks the geometry first. Every request becomes a ticket that the printer owner moves through Requested, Accepted, Printing, and Delivery, or declines with a reason, and requesters can withdraw their own print before it reaches the bed. Slicer specifics like layer height, infill, supports, and temperatures ride along as notes on the ticket, a low, medium, or high priority keeps the owner's queue ordered, and a conversation thread on each ticket replaces lost chat threads. The owner curates the materials and colours on the shelf, with swatches and a rainbow whatever option, and defines the tips requesters can offer, preferred ones starred. Finished jobs collect in a filterable history: Print again reopens the form with the model copied server-side and fields adjustable, and one click sends a model straight to PrusaSlicer. Accounts are invitation-only; access can be revoked with a global sign-out, every change lands in an audit trail that cannot be edited, and the same tickets are exposed over a JSON API documented in OpenAPI 3.1 with a Swagger console. 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.

Inference Gateway puts a single OpenAI-compatible API endpoint in front of OpenAI, Anthropic, Groq, Cohere, Ollama, DeepSeek, Google, Mistral, MiniMax, Moonshot, Nvidia, and llama.cpp, so your application code never changes when you switch models or providers. The Go binary starts on port 8080 and normalizes authentication, streaming protocols, and response formats across all backends transparently. Native Model Context Protocol support auto-discovers tools from connected MCP servers and injects them into LLM requests without client-side management, enabling server-side tool execution across any provider that supports function calling. Agent-to-Agent protocol integration allows distributed agent communication through a declarative Agent Definition Language that generates production-ready Go or Rust servers from a single YAML manifest. The dedicated Kubernetes Operator manages Gateway, Agent, MCP, and Orchestrator custom resources with automatic HPA scaling, OIDC authentication, and service discovery that rebuilds MCP configurations when the discovered server set changes. Prometheus metrics and OpenTelemetry tracing provide full request-level observability across the entire inference pipeline. Middleware controls enable per-request provider selection, model routing, and fallback strategies. Official SDKs in Go, Python, TypeScript, and Rust provide typed client interfaces with streaming support. Docker Compose deployment requires only environment variables for API keys. A CNCF Sandbox applicant. 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.

Stop handing your daily food diary, body weight history, and biometric sleep data to ad-funded fitness clouds: NutriTrace brings complete calorie, macro, micronutrient, and wellness tracking to your own server with an offline-capable web PWA and native Android app. Log meals across customizable slots using live Open Food Facts and USDA FoodData Central lookups, barcode scanning, AI nutrition label OCR, or direct recipe imports from Mealie and CookTrace. Calorie and macro targets adjust automatically through three goal engines, including an inspectable Adaptive TDEE algorithm that regresses 35 days of intake against scale weight at 7,700 kcal per kilogram. Connect Fitbit via Google Health, Garmin, Withings, and Android Health Connect to pull steps, heart rate variability, sleep stages, and body composition into unified Sleep, Readiness, and Resilience scores. An integrated Trace AI assistant and Smart Log let you log meals by voice, text, or meal photo, while the opt-in Model Context Protocol server exposes 20 read, write, and management tools to Claude Desktop, Cursor, and Codex. Multi-user households get OpenID Connect SSO and one-step migration from MyFitnessPal, Cronometer, Lose It, and Waistline. 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.

Your next product listing does not need a photo studio, because EcomGen turns the product facts and photos you already have into a complete image suite you approve shot by shot. Each product becomes a project that remembers its verified facts, banned claims, brand rules, target market, and platform requirements, so generation starts from what is true about the item rather than a fresh guess every time. An agent reads the template library and drafts a storyboard with a role for every image: hero, pain point, comparison, scene, detail, trust, variant, and conversion, and you edit and confirm each final prompt before anything renders. Twenty-five built-in templates and more than 1,100 reusable suites across 24 top-level categories keep a catalog visually consistent, and the Suite Forge can reverse-engineer a winning competitor set into a reusable template of your own. Product-truth pixel protection keeps the real item intact instead of letting the model improvise a different product. Finished shots can be masked, repainted, outpainted, and split into editable PNG and PSD layers, reviewed one by one, then exported as a ZIP with a manifest, while the model library keeps virtual models consistent across poses and scenes. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Pictaria Server sits beside your Immich instance and makes a photo library of thirty thousand images legible again, surfacing the trips, people, and best shots buried under duplicate bursts and forgotten folders. It starts with Insights: a clickable timeline of your years, detected trips, a people constellation linked by who appears together, place and camera rankings, and records like your busiest day. Click any number and you land on the actual photos, ready to open in Immich, turn into an album, or send onward. Enrich puts an AI to work describing what it sees, writing captions, menu-driven tags, and quality scores for every photo using a local model through Ollama or LM Studio, or a cloud provider with your own key, on a scheduled daily budget. Those tags can sync back into Immich, so searching improves everywhere, not just here. Curate then collapses bursts and near-duplicates into stacks with a suggested keeper, and one keystroke keeps the best and skips the rest. Smart Albums compose rules from people, places, dates, cameras, and tags, preview the result count, and maintain real Immich albums on a schedule. Frame owners get remote control, display metrics, and voice photo questions, while the Activity feed and runtime-editable Settings keep the whole box manageable from the browser. It connects with a limited-permissions API key and cannot delete a single photo. 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.