Run a Stack Overflow of your own: Apache Answer brings the question-and-answer format in-house, maintained under the Apache Software Foundation with an Apache-2.0 license. You use it to run a community forum, product help center, or internal knowledge base where content lives as questions and answers rather than wiki pages. It ships the mechanics that make that format work: voting, accepted answers, a reputation system with privilege levels, tagging, full-text search with filters, revision history on every edit, and admin/moderator/user roles. Content is written in Markdown with real-time preview and code syntax highlighting. A plugin system covers OAuth login (Google, GitHub), S3 storage, external search backends like Algolia, and Akismet anti-spam, and a REST API exposes platform data for integration. The backend is Go, the frontend React, and it runs against SQLite, MySQL, or PostgreSQL. Self-hosting replaces per-seat tools like Stack Overflow for Teams with a flat-cost instance where you own all the content.

Trello's board model on your own server: Planka is an open-source Kanban project management tool. Boards organize into projects with lists, cards, labels, due dates, checklists, file attachments, and per-card stopwatch time tracking, all managed through drag-and-drop. Updates propagate over WebSockets, so a teammate moving a card or adding a comment appears instantly for everyone without a refresh - a genuine differentiator among self-hosted boards. Card descriptions use a full Markdown editor, custom fields adapt cards to your workflow, and views switch between Kanban, grid, and list layouts. Authentication supports OpenID Connect single sign-on with Google, Azure AD, Okta, or any OIDC provider - a feature Trello reserves for enterprise plans - and notifications reach 100+ channels including Slack, Discord, Telegram, and SMTP via Apprise. A REST API with 50+ webhook events supports custom integrations, and one-click board import eases migration. Built with React and Node.js on PostgreSQL, translated into 35+ languages, deployed via Docker.

An entire backend in a single Go executable: PocketBase embeds SQLite with realtime subscriptions, authentication and user management, file storage, and an admin dashboard, all behind a REST-ish API. SQLite runs in WAL mode, which outperforms client-server databases for the read-heavy workloads typical of small and mid-sized apps. Authentication supports email/password, one-time passwords, and 15+ OAuth2 providers including Google, Apple, and GitHub, with stateless tokens. Clients subscribe to record changes over server-sent events, and official JavaScript and Dart SDKs cover web, mobile, and Flutter frontends. Collections, rules, and API access permissions are managed visually in the admin UI. When you need custom logic, extend it with JavaScript hooks running in the embedded JS VM of the prebuilt binary, or import PocketBase as a Go library and compile custom business logic into your own single-file backend. File storage attaches uploads to records with thumbnail generation for images and optional S3-compatible external storage. All state lives in one pb_data directory, so backup is a directory copy and upgrade is replacing a binary - one of the lowest-maintenance backends you can run. The contrast with Firebase is the point: where usage-based pricing scales with reads, writes, and bandwidth, PocketBase runs the entire backend at flat hosting cost, and the data is a plain SQLite file you can copy anywhere. MIT-licensed.

A question goes in; a cited, long-form report comes out - GPT Researcher is an open-source autonomous research agent. A planner agent decomposes the query into sub-questions, execution agents crawl 20+ web sources in parallel with JavaScript-enabled scraping, and a publisher aggregates findings into a 2,000+ word report with inline citations, exportable to PDF, Word, and Markdown. The Deep Research mode extends this recursively: each result yields follow-up questions that are explored to configurable breadth and depth in a tree pattern, while accumulated learnings, citations, and visited URLs are shared across branches. It also researches local documents (PDF, CSV, Word) alongside the web. LLM and search providers are pluggable, including OpenAI, Anthropic, Google, DeepSeek, and Ollama for models, and Tavily, Google, Bing, DuckDuckGo, and SearXNG for retrieval. It ships as a Python package, a FastAPI server with web frontend, a Docker image, and an MCP server for use inside Claude or Cursor. MIT-licensed.

Salesforce's core workflow, open-source and on your own server: Twenty is a modern CRM built as exactly that alternative. It ships the standard CRM objects out of the box: people, companies, opportunities, notes, and tasks, displayed in table and kanban views with drag-and-drop and real-time updates. Its defining technical feature is a metadata-driven data model: you define custom objects and fields in the UI, and the backend regenerates its GraphQL schema at runtime, so a new object gets working queries, mutations, filters, and sorting within seconds, with no migrations to run - adapting the CRM to your sales process never requires code changes. A REST API is auto-generated from the same schema, GraphQL subscriptions push real-time updates, and webhooks fire on record changes for external integration. A visual workflow builder automates actions like notifications and field updates, TypeScript-based apps extend the platform with custom logic and frontend components, and email and calendar sync pulls Gmail messages and meetings onto contact timelines so communication history sits next to the record. The stack is NestJS with TypeORM, PostgreSQL, Redis, and BullMQ on the backend, React with Jotai on the frontend. Self-hosting on RepoCloud means unlimited users with no per-seat licensing - the pricing model that penalizes growing teams on commercial CRMs - and your pipeline, contacts, and deal history live in your own PostgreSQL database rather than a vendor's.

Describe a component in natural language and watch it render: OpenUI, from Weights & Biases, is an open alternative to Vercel's v0. Type a prompt like "a dark-themed dashboard with a sidebar and charts" and the LLM renders working HTML with Tailwind styling live in the browser. You then iterate conversationally, asking for changes until the design is right, and convert the result to React, Svelte, or Web Components for use in a real project. The backend is Python with LiteLLM routing, so it works with OpenAI, Anthropic, Gemini, Groq, and Mistral API keys, or fully offline against local Ollama models, including vision models like LLaVA that can generate UI from screenshot input - feed a screenshot and the model reproduces or riffs on an existing interface. Generated markup is inspectable at any point, with light and dark mode toggles, theme selection, and responsive previews across device sizes. The practical effect is compressing the mockup-review-revise loop from hours to minutes: a described layout renders in seconds and iterates through follow-up prompts, and because output converts to real framework code, prototypes feed directly into production codebases instead of staying trapped in a design tool. Self-hosting keeps unreleased product interfaces and prompts on your own server, and LiteLLM routing lets you pick the model per task - a cheap fast model for rough drafts, a stronger one for final passes, or free local models for unlimited experimentation.

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.

Every major model provider behind one ChatGPT-style interface: LibreChat spans OpenAI, Anthropic, Google, Azure, AWS Bedrock, Vertex AI, Groq, Mistral, OpenRouter, DeepSeek, and any OpenAI-compatible endpoint including local Ollama. You can switch models mid-conversation and compare providers without changing tools. Its Agents framework builds no-code custom assistants with tool access via Model Context Protocol servers, file search over uploaded documents through an optional pgvector-backed RAG service, and a sandboxed Code Interpreter that executes Python, JavaScript, Go, C++, Java, PHP, and Rust. Artifacts render React components, HTML, and Mermaid diagrams directly in chat, and image generation works through DALL-E and other configured providers. Multi-user support is enterprise-grade, with OAuth, SAML, LDAP, and two-factor authentication, per-user conversation history in MongoDB, and Meilisearch-powered search across all messages and files, plus reusable presets, forkable threads, and persistent memory across conversations. The economics favor teams: instead of a ChatGPT Plus seat per person, everyone shares one instance billed per API token, with access to every provider rather than one - and providers see individual API calls, not your accumulated organizational knowledge. Deployment is Docker Compose; API keys and endpoints are configured through .env and librechat.yaml.

Prototype multi-agent AI systems without writing orchestration code: AutoGen Studio is Microsoft's low-code interface over the AutoGen AgentChat framework. You compose teams of LLM-powered agents in a visual Team Builder, either by drag-and-drop from a component library or by editing the declarative JSON specification directly. Each agent gets a model, a prompt, tools (Python functions), and the team gets termination conditions and an orchestration pattern, sequential or LLM-driven. The Playground runs teams interactively with live message streaming between agents, a visual control-transition graph, tool-call and code-execution tracking, and pause/stop controls, which makes it a practical debugger for agent behavior. Finished teams export as JSON for use in any Python application via the TeamManager class, or serve as an API endpoint. Any OpenAI-compatible model endpoint works, including local servers like Ollama or vLLM. Microsoft labels it a research prototype: use it for prototyping and evaluation, and build production systems on the underlying AutoGen framework.

The most widely deployed open-source BI tool, Metabase is a visualization and query layer that sits on top of your existing databases without ingesting or copying data. Non-technical users ask questions through a visual query builder with drill-through menus that answer follow-ups like "broken down by month" without writing a new query, while analysts use the native SQL editor with variables and templates for complex work. Questions assemble into interactive dashboards with filters, auto-refresh, fullscreen mode, and custom click behavior, and dashboard subscriptions email or Slack scheduled reports to stakeholders. It connects to 20+ data sources including PostgreSQL, MySQL, MongoDB, SQL Server, BigQuery, Snowflake, Redshift, and ClickHouse - always querying in place, so there is no second data store to secure, sync, or pay for, and results are always current. Models and metrics let a data team define official, reusable starting points so self-service stays consistent, collections with permissions organize content, and alerts fire when a metric crosses a threshold. The practical effect is cutting the ad-hoc query queue that lands on the data team, since non-technical staff can answer their own questions. Written in Clojure, licensed AGPL, and shipped as a single JAR or Docker image with an embedded application database - a working BI instance runs before most tools finish their installer - the open-source edition has no limits on users, dashboards, or connected databases, where commercial BI platforms price per viewer as well as per creator.

A campaign engine wrapped around a contact database: Mautic, the largest open-source marketing automation platform, replaces HubSpot or Marketo without per-contact pricing. Contacts arrive through forms, landing pages, imports, or the REST API and flow into segments: dynamic filters that update automatically from behavior, custom profile fields, or point scores. Segments decide who qualifies; campaigns decide what happens. The drag-and-drop Campaign Builder composes multi-step workflows from actions, positive/negative decision trees, and conditions (field values, tags, device type, segment membership, point thresholds), with static or relative delays, a Jump-to-Step action for moving contacts between branches, and handoffs that push contacts into CRMs or entirely different campaigns. Messaging covers email, SMS, and web/app push out of the box, with A/B testing and a drag-and-drop email builder; dynamic website content swaps page sections per known contact. Lead scoring assigns points for clicks and visits with decay for inactivity, while stages track funnel position. Native integrations cover Salesforce, HubSpot, Zoho, and Dynamics, plus a full REST API for custom sync. It runs on PHP and MySQL with cron jobs processing campaigns and segment rebuilds - self-hosting keeps your entire contact database and behavioral history under your control.

Quote to cash in one web application - create quotes, convert them to invoices, record payments, track customers: iDURAR is an open-source ERP and CRM platform for small and medium-sized businesses. Built on the MERN stack (MongoDB, Express, React, Node.js) with Ant Design components and Redux state management, it presents a clean SaaS-style interface that needs little onboarding. Core modules cover invoice management with PDF generation and email delivery, payment recording against invoices, quote and proforma handling, customer records, and accounting views over the resulting data. Multi-currency support and localization make it usable for internationally operating teams. Because the whole stack is JavaScript with an API-first backend, extending it - custom fields, new modules, integrations - is approachable for any Node/React developer rather than requiring a specialist ERP skill set. Deployment is straightforward via Docker with a MongoDB instance. Licensed under AGPL-3.0 with free commercial use; a hosted enterprise version exists but the self-hosted edition is fully functional.

Your Gmail, Google Workspace, or Outlook inbox, worked by an AI assistant: Inbox Zero sits on top of the account you already have. Its core idea is rules written in plain English - tell the assistant "label invoices and file the PDF to Drive" or "archive cold outreach unless they mention my company" - and it executes against every incoming message. Emails that need a response arrive with a pre-drafted reply written in your tone, learned from your email history and calendar context. Reply Zero tracks what you owe responses to and what you're waiting on; the Bulk Unsubscriber surfaces newsletters you never read (with read-rate analytics) for one-click unsubscribe-and-archive; the Cold Email Blocker auto-archives unsolicited pitches based on your own definition of "cold." Smart Filing routes attachments - receipts, contracts, PDFs - into the right Google Drive or OneDrive folder, and Slack/Telegram integration lets you read, draft, and triage without opening a mail client. Email analytics show top senders and volume trends. It is not a new email client: everything happens in your real mailbox using native filters. Self-hosting means your mail content and the LLM calls that process it run on infrastructure you control.

Prompt, run, edit, and deploy full-stack Node.js applications from a browser tab: Bolt.diy is the official open-source version of Bolt.new's AI coding agent. Its foundation is StackBlitz's WebContainer technology - a sandboxed in-browser Node.js environment where the AI controls the whole stack: filesystem, npm, dev servers, terminal, and browser console. That means the agent does not just generate code; it installs dependencies, runs Vite or Next.js, reads errors, and fixes them. The defining difference from Bolt.new is model choice per prompt: 19+ providers including OpenAI, Anthropic, Gemini, DeepSeek, Groq, Mistral, Amazon Bedrock, and local models via Ollama or LMStudio, extensible through the Vercel AI SDK. Development ergonomics include live preview, a diff view of AI changes, codebase search, file locking to prevent generation conflicts, 15+ starter templates (React, Vue, Next.js, Astro, Svelte, Expo), and MCP support for external tools. Projects integrate with Git and Supabase, and deploy in one click to Vercel, Netlify, or GitHub Pages.

Your users belong in your own database - Authorizer, an open-source authentication and authorization server shipping as a single Go binary, keeps them there. It connects to 13+ backends - PostgreSQL, MySQL, SQLite, SQL Server, MariaDB, MongoDB, Cassandra, ScyllaDB, ArangoDB, DynamoDB, Couchbase, YugabyteDB, PlanetScale, and CockroachDB - so identity data lives beside the application it protects instead of in an auth vendor's cloud. The server is fully OAuth 2.0 and OpenID Connect compliant, including authorization code flow with PKCE, a JWKS endpoint, token revocation, and nine JWT signing algorithms. Login options cover email/password, magic links, TOTP multi-factor, SMS OTP via Twilio, and social providers including Google, GitHub, Apple, Microsoft, and Discord. Authorization goes beyond roles: an embedded OpenFGA engine provides Zanzibar-style relationship-based permission checks in process. APIs are exposed over GraphQL, REST, and gRPC, with SDKs for JavaScript, React, Go, and Python, plus themeable built-in login pages and an admin dashboard. Apache 2.0 licensed.

Typeform's conversational format, self-hosted: HeyForm is the open-source form builder that presents one question at a time. Forms present one question at a time, which measurably improves completion rates compared to long static pages. It supports 40+ field types, from text, email, and phone inputs to picture choices, date pickers, star ratings, signatures, and file uploads. Conditional logic shows or hides questions based on earlier answers, routes respondents to different endings, and redirects to URLs, so a single form can serve multiple flows. Completed submissions land in a results dashboard with drop-off and completion analytics, and connect outward through webhooks or integrations with Zapier, Make, Google Sheets, Notion, Airtable, and Slack. Theming covers fonts, colors, backgrounds, and custom CSS, so embedded forms look native to your site rather than like a third-party widget; the JavaScript embed library renders them inline, as popups, or full-page, with shareable standalone links as the default. Team workspaces and projects with member management let multiple teams share one instance without mixing data. Self-hosting removes per-response pricing entirely - unlimited forms and submissions for flat hosting cost - and keeps lead data, feedback, and quiz answers in your own MongoDB, simplifying GDPR compliance. The stack is a NestJS server and React webapp backed by MongoDB and KeyDB, distributed under GPLv3 as a Docker image.

Live web search plus an LLM of your choice: Farfalle is an open-source, self-hosted answer engine in the Perplexity mold. Queries route through one of several search providers - self-hosted SearXNG for a fully independent stack, or Tavily, Serper, and Bing APIs - and the model composes a cited answer from the retrieved results. Model flexibility is the core design: run llama3, mistral, gemma, or phi3 locally through Ollama for zero per-query cost and full privacy, use cloud models like GPT-4o or Groq-hosted Llama 3 for speed, or route to any provider via LiteLLM. An Expert Search mode uses an agent that plans a multi-step search strategy and executes it for harder questions, and chat history keeps prior research sessions available. The stack is a Next.js and shadcn/ui frontend over a FastAPI backend with Redis rate limiting, shipped as a pre-built Docker image. A browser search-engine entry pointing at your instance makes it the default search from the address bar. Paired with SearXNG and Ollama, the whole pipeline runs with no external API at all.

Roughly $1M of development work, open-sourced: Maybe Finance began as a $249/year commercial personal finance product before the company released it all. It aggregates bank accounts, credit cards, loans, investments, crypto, and real estate into a single net worth dashboard with historical trend charts - replacing the spreadsheet that usually glues a whole portfolio together. Transactions are categorized and tagged with rules, with merchant tracking and search across imported or synced activity; budgets track spending by category against plan; and the investment view follows holdings, cost basis, and returns across brokerage accounts. Multi-currency support converts accounts held in different currencies into a single reporting currency, bank synchronization works through Plaid where supported, and manual CSV import covers any institution. An optional AI assistant answers questions grounded in your own financial data. Because the app was built as a paid product with professional design before being open-sourced, its interface quality exceeds most community finance tools - and self-hosting means your balances and transactions are not monetized by a free app or gated behind an annual subscription. The stack is Ruby on Rails with Hotwire on PostgreSQL, licensed AGPL-3.0 and deployed via Docker. The original repository is archived; development continues in the community fork Sure, compatible with the same self-hosted setup.