Cloudron Forum

@scooke There is a relatively new (and strange) restriction on "open source" community edition rocketchat: only 25 users. Sessions are restricted, too.

I was sure that Cloudron had started supporting Zulip. I looked for it and was surprised that it is still not there.

Improvements have been considerable since the original request. Lets support Zulip!

Based on the search results, here are the main advances in Zulip since March 2018 categorized by client and server improvements:

Server Improvements

Performance & Scalability (2018-2025)

• Reworked message feed rendering with improved client-side caching for instant rendering of recently visited views
• Optimized presence synchronization protocol using dramatically less network/CPU resources in large organizations
• Squashed almost 700 database migrations to optimize new database creation runtime
• Improved performance of creating channels with thousands of initial subscribers
• Replaced pip with uv for better dependency management
• Implemented typed_endpoint abstraction backed by Pydantic v2 for improved performance and readability

Security & Privacy (2020-2025)

• Added end-to-end encryption (E2EE) for mobile push notifications
• Hardened message-fetching code path against access control bugs
• Added server-to-server Zoom integration with better security
• Improved LDAP integration with better user role synchronization
• Added data export features with member consent controls
• Fixed multiple CVEs including authentication bypass and channel permission issues

File Management (2021-2025)

• Added TUS chunked-upload protocol support for arbitrarily large files
• Increased default MAX_FILE_UPLOAD_SIZE from 25MB to 100MB+
• Added support for HEIC and TIFF image thumbnails
• Improved video previews with better play button experience

Administration & Management (2019-2025)

• Channel archiving now preserves names and content access (with unarchiving support)
• Group-based permissions system with granular channel controls
• SCIM integration for user synchronization
• Custom welcome messages for new users
• Automated mobile push notification service registration
• Improved user role management and permission systems

Client Improvements

User Interface (2020-2025)

• Channel folders for organizing team channels
• Recent topics view for channel overview
• Improved left sidebar with search filters and unread count controls
• Cleaner image/video thumbnail layouts
• Enhanced message composition with better destination visualization
• Modernized settings UI with minimal, clean design

Messaging Experience (2019-2025)

• Message reminders functionality
• Support for channels without topics
• Bulk message management (mark as read for muted topics)
• Improved search with better operators and filters
• Pasting large text now convertible to uploaded files
• Read receipts and user status/availability

Mobile App (2021-2025)

• Complete rewrite with modern architecture
• Added search, status, read receipts, and sharing features
• End-to-end encrypted push notifications
• Autocomplete for user group mentions
• Broader TeX math support

Integration Ecosystem (2018-2025)

• Added integrations with OpenProject, OpenSearch, Airbyte, Onyx
• Improved GitHub, GitLab, Jotform, Linear, NewRelic integrations
• Removed deprecated integrations (OpsBeat, Desk.com)
• Enhanced SAML and Microsoft Entra ID authentication

The platform has evolved from a basic chat system to a comprehensive team collaboration platform with enterprise-grade security, scalability, and user experience improvements across all clients.

Ever hopeful, we thought it would be fun to try and one-shot a cloudron package of anythingllm using Grok.

We created a repo on cloudron's demo site (slick looking new GUI, by the way) using Gogs.

Now we want to find the install from custom source but can't see how. Could anybody show how to do it / Screenshot?

AnythingLLM has had enormous improvements since it was first requested by @TheMoodBoardz
In particular, support for agents and RAG has been advanced to try and keep up with the huge demand in this area.

Installing AnythingLLM on GNU+Linux now automatically installs Ollama, too.

Here is a summary of the main advances. If you want to make RAG and ai agents easier, AnythingLLM would be a great application to support.

Comparison between v1.0Based.0 and v1.9.0 of AnythingLLM, here are the most significant changes categorized into key areas:

Agent System Overhaul

• Complete redesign of the agent experience with real-time streaming tool calls
• Agents can now download and ingest web files (PDF, Excel, CSV) during conversations
• All providers and models support agentic streaming capabilities

Local LLM Integration

• Added Microsoft Foundry Local integration for Windows/MacOS (free local LLM solution)
• Linux now includes Ollama (0.11.4) built-in for immediate local LLM support
• ARM64 support for both Linux and Windows platforms

Platform & Infrastructure

• Expanded Linux support with automatic apparmor rules and .desktop file creation
• NVIDIA NIM being phased out in favor of newer solutions
• Upgraded core Electron version for improved stability

UI/UX Improvements

• Model swap capability during chats (Ctrl/Cmd+Shift+L)
• Workspace and thread searching in sidebar
• System prompt version tracking and restoration
• Enhanced mobile support with Android beta app

Developer Features

• MCP (Model Context Protocol) compatibility
• No-code AI agent builder
• Enhanced API endpoints and developer tools
• Multi-modal support for both open and closed-source LLMs

Internationalization

• Added Portuguese and Estonian translations
• Improved multi-language OCR support
• Expanded localization coverage

The update represents a massive leap from the initial version, transforming from a basic document chat system to a comprehensive AI agent platform with robust local operation capabilities and extensive customization options.

thanks to @timconsidine, we have facilmaps available on Cloudron, which is fantastic, but lets get our 70TB monster zoom maps going!

https://github.com/public-transport/transitous/

This project might be useful with some others Cloudron supports.

What is the latest news on official support for BTCpayserver on Cloudron?

@jdaviescoates made a great request here and very early in the BookWyrm's development. A lot has improved since then. Here is an ai generated summary. Lets support BookWyrm on Cloudron!

Bookwyrm Improvements: v0.0.1 to v0.8.0

Core Features

• Introduced user export and import functionality, with files now saved to /exports in local storage by default, and automatic deletion of import files post-completion.
• Updated minimum Postgres version to 14 (17 preferred) and Django to 5.2 for enhanced core stability and compatibility.
• Added management command to merge works, improving data handling for book entries.
• Added dates to RSS feeds and sorted them by most recent first for better content delivery.

User Interface/UX

• Display "no rating" instead of blank stars for reviews without ratings, improving visual clarity.
• Use translated shelf names in notifications for shelving other editions, enhancing multilingual support.
• Retain subjects and authors in the new book form even if validation fails, reducing user frustration during input.

Social/Community Features

• Show all reports (not just open ones) when linked from user admin, aiding moderation.
• Fixed federation issues with GoToSocial and inconsistent KeyId in headers, improving interoperability.

Book Management

• Optimized get_audience function by fetching only IDs, streamlining book-related queries.
• Enabled merging of works via a new management command for cleaner book database management.

Performance/Optimization

• Optimized get_audience by only fetching IDs to reduce query overhead.
• Increased default file upload limit to 100MiB, configurable via MAX_UPLOAD_MiB in .env.

Bug Fixes

• Fixed federation with GoToSocial and inconsistent KeyId in headers.
• Retained subjects and authors when new book form fails validation.
• Updated nginx config requirements, including new location blocks for static files and access restrictions to /images/.

Security/Privacy

• Included important security fixes in v0.8.0, with warnings to manually delete publicly available user export files (.gz) from images directories and S3 storage.
• Changed default DEBUG to false for production security; assumed HTTPS unless DOMAIN=localhost.
• Added optional S3 support for export files via new env variables like USE_S3_FOR_EXPORTS=true, with separate credentials for isolation.
• Deprecated USE_HTTPS and enforced HTTPS in production to prevent HTTP URLs in user/object IDs.
• Added nginx config to block access to non-image files in /images/ with 403 responses.

Documentation/Setup

• Introduced NGINX_SETUP env variable (values: reverse-proxy or https, defaults to https) for better proxy and dev environment handling.
• Updated default PORT to 443 (or 80 for localhost) if unset, differing from previous 1333 default.
• Provided detailed upgrade instructions for all environments, including Docker steps like docker compose down, database version upgrade to Postgres 17, and rebuilding with Django 5.2.
• Required manual nginx config updates in v0.7.2 for static file handling and security blocks.

Other

• Enabled optional S3 storage for exports with new env variables like EXPORTS_S3_CUSTOM_DOMAIN and EXPORTS_STORAGE_BUCKET_NAME.
• Encouraged deletion of legacy export files in S3, especially for Backblaze (B2) or Cloudflare (R2) due to potential unsecured buckets.

It doesn't use code from any of the existing web-browsers.

https://ladybird.org/

I tried creating the above post using AI. The formatting wasn't beautiful, so I retried with AI, asking it to improve the formatting. It didn't get much better.

I shall tidy it up later but am leaving it here for a while so you can see how it looks.

Main Page: https://github.com/HenryPenton/train_dashboard
*

• Git: https://github.com/HenryPenton/train_dashboard
• Licence: MIT
• Dockerfile: Yes (pre-built Docker images available)
• Demo: N/A

• ---

• Summary: A full-stack application for live train and tube status updates, route planning, and departure information. Built with Next.js for the frontend and FastAPI for the backend, it integrates with Transport for London (TFL) and Real Time Trains APIs to provide real-time data.

• ---

• Notes: I like this app because it provides a self-hosted, customizable dashboard for monitoring UK public transport, ideal for commuters or enthusiasts who want to avoid proprietary apps. The retro-inspired UI is a nice touch, and it's responsive for use on various devices. Concerns include it being under heavy development, potential bugs, and the need for API credentials from Real Time Trains for full functionality.

• ---

• Alternative to / Libhunt link: No direct self-hosted alternatives found, but related transportation projects at https://www.libhunt.com/topic/transportation (e.g., https://selfhosted.libhunt.com/trs-alternatives for CLI timetables)
• Screenshots:
• 
• :

https://immich.app/blog/google-flags-immich-as-dangerous

@marcusquinn made a great application request here. The Qdrant project has become increasingly useful for AI and a lot has happened since the initial request to make the project even better. Here is an ai generated summary:

Summary of New Features and Improvements in Qdrant from v1.1.0 to v1.15.5

Based on the changelog from this GitHub comparison, here's a categorized breakdown of the major updates. Early versions (1.1–1.2) laid the foundation with vector search and HNSW indexing, while later releases added scalability, hybrid search, and performance tweaks. This evolved Qdrant into a robust vector database for AI applications.

Vector Search and AI Features

• Hybrid Search Support: Introduced sparse vectors alongside dense ones (e.g., BM25/TF-IDF generalization for keyword matching in transformer models, starting ~1.3–1.4).
• Recommendation Strategies: New options like sum_scores for relevance feedback and server-side score boosting with custom formulas (v1.14.x).
• Local BM25 Inference: Added built-in BM25 scoring (v1.15.x), cutting reliance on external models for text ranking.

Performance and Optimization

• HNSW Graph Improvements: Incremental building on segment merges (v1.14.x), parallel batch searches, and memory optimizations.
• Memory/Disk Management: Better disk cache eviction (v1.14.x), adjustable log buffers, and vector quantization reducing RAM usage by ~97% (since ~1.3).
• SIMD Acceleration: x86-64 and Neon support for faster vector computations (ongoing from ~1.5).
• Async I/O and WAL: Write-ahead logging for data durability and io_uring for max NAS throughput (major in 1.5–1.10).

Storage, Sharding, and Scalability

• Distributed Deployment: Full horizontal scaling with sharding/replication and zero-downtime updates (matured in 1.10–1.15).
• Mutable Map Index: Boosted performance for full-text, integer, and other indexes (v1.15.x).
• On-Disk Storage: Dynamic on-disk vectors with cache controls (replaced deprecated mmap_threshold in 1.15.x).
• Resharding/Recovery: Improved shard recovery against crashes and aborted resharding for dead replicas (v1.14.x).

Security and Access Control

• JWT/Payload Filters: JWT for read-only access and payload-based filters (deprecated in 1.15; now collection-level ACLs).
• Access Controls: Shift to secure, collection-based permissions (recommended since 1.15).

Web UI and User Experience

• Enhancements: Query auto-completion (v1.14.x) and shard distribution in cluster views (v1.15.x). Debian packages now include web UI (v1.15.x).

Payload and Filtering

• Advanced Filtering: JSON payloads, keyword/full-text matching, numerical/geo ranges, and complex conditions (e.g., should, must, must_not). Optimized query planning with payload indexes (steady improvements since ~1.2).

Bug Fixes and Consistency

• Key fixes: Delayed flush wrappers prevent lost index updates (1.14.x); storage issues for nulls/booleans deferred until flush (1.15.x); linear decay range bounds corrections (1.15.x).
• Others: Scroll locks, segment mods, date-time errors in formulas, and corrupt point eviction (across 1.14–1.15).

Deprecations and Removals

• Removed init_from for consistency (use migrations/snapshots; 1.15.x).
• Deprecated max_optimization_threads (1.15.x), Lock API, and payload-based JWT filters.
• Deprecated mmap_threshold in favor of on_disk (1.15.x).

Overall, Qdrant went from basic vector search to an enterprise-grade tool emphasizing hybrid search, reliability, and AI integration. Full details in the GitHub link. Let me know if you want more on a specific version or category!

• Main Page: https://cobalt.tools/
• Git: https://github.com/imputnet/cobalt
• Licence: AGPL v3.0
• Dockerfile: Install: Use cobalt.tools online; or docker run -p 3501:3501 cobalt/cobalt for self-host.
• ? https://github.com/imputnet/cobalt/issues/722
• Demo: https://cobalt.tools/

---

• Summary: Download videos and watch them offline instead of viewing them on proprietary platforms.

---

• Notes: large development effort, it works and does so across several platfroms. Docker support might be tricky

---

• Alternative to / Libhunt link: eg invidious, yt-dlg, yt-dlp, https://tartube.sourceforge.io/ https://www.libhunt.com/r/cobalt
• Screenshots:
  
  

@nebulon Thanks, we ran into this issue, too. What is it in Cloudron 9 that immich requires?

@marcusquinn Thanks for that. Looks like my post is a duplicate of:
https://forum.cloudron.io/topic/13141/litellm-openrouter-self-hosted-alternative-proxy-provides-access-to-openai-bedrock-anthropic-gemini-etc?_=1759700788590

(I posted 3 apps, checked the first one, and then forgot to check if the other two had already been requested!)

@girish What is the procedure to merge this dupe with @jagan 's earlier request here?

https://forum.cloudron.io/topic/13141/litellm-openrouter-self-hosted-alternative-proxy-provides-access-to-openai-bedrock-anthropic-gemini-etc?_=1759700788590

I can see the topic tools, merge option, but it is a bit confusing on how to paste/include jagan's thread.

@timconsidine SillyTavern is good for multi-modal integration, like text-to-speech, speech-to-text, image generation and for making sustaining context a bit easier for drama/RPGs.

Since you are curious, some of the other app requests I recently made are there so that cloudron would support some of the tools most frequently used by ai, according to OpenRouter statistics.

Here is how ai (GPT5) explains it:

Short answer: SillyTavern and LibreChat target different use cases. Having LibreChat doesn’t make SillyTavern redundant.

Why someone would pick SillyTavern
Character/roleplay first: Built-in support for character cards, group chats/party mode, world info/lorebooks, long-form memory and style presets tailored for RP and creative writing.
Strong local-LLM focus: Connects directly to many non–OpenAI-compatible backends (e.g., KoboldCpp, text-generation-webui, llama.cpp) as well as OpenAI/Anthropic/Ollama/OpenRouter. Great for people running purely local models.
Immersive extras: Optional “SillyTavern-Extras” for TTS/STT, translation, and image generation integrations for story/RP flows.
Lightweight and personal: Mostly a client/front-end; simple to run, low overhead, well-suited to single-user creative use.

Where LibreChat shines
General-purpose, ChatGPT-like UX: Good for Q&A, productivity, and teams.
Multi-user and org features: Auth, sharing, and workspace-style flows.
Standards-based backends: Primarily OpenAI-compatible APIs (also works with Ollama and other compatible gateways).
Tools/RAG/file workflows: Better fit for business/knowledge tasks and agents.

• Main Page: https://www.newapi.ai/en/
• Git: https://github.com/QuantumNous/new-api
• Licence: Apache 2 / Dual-Use https://github.com/QuantumNous/new-api?tab=License-1-ov-file#readme
• Dockerfile: Yes https://github.com/QuantumNous/new-api/blob/main/Dockerfile
• Demo: link

---

• Summary: next-generation large model gateway and AI asset management system, based on One API for secondary development. The project aims to provide a unified interface for managing and using various AI model services, including but not limited to OpenAI, Anthropic, Midjourney, and more.

---

• Notes: why you like it/concerns

---

• Alternative to / Libhunt link: e.g. https://selfhosted.libhunt.com/umap-alternatives
• Screenshots: images, brand logo

• Main Page: https://www.litellm.ai/
• Git: https://github.com/BerriAI/litellm
• Licence: MIT? https://github.com/BerriAI/litellm?tab=License-1-ov-file#readme
• Dockerfile: Yes https://github.com/QuantumNous/new-api/blob/main/Dockerfile
• Demo: Youtube video on main page

---

• Summary:
• Call 100+ LLMs using the OpenAI Input/Output Format
• Translate inputs to provider's completion, embedding, and image_generation endpoints
• (Consistent output, text responses will always be available at ['choices'][0]['message']['content']
• Retry/fallback logic across multiple deployments (e.g. Azure/OpenAI) - Router
• Track spend & set budgets per project LiteLLM Proxy Server

---

• Notes: Supported LLM providers:
  https://docs.litellm.ai/docs/providers

---

• Alternative to / Libhunt link: OpenRouter
  https://alternativeto.net/software/litellm/

• Screenshots: images, brand logo

• Main Page: https://sillytavern.app/
• Git: https://github.com/SillyTavern/SillyTavern
• Licence: AGPL-3.0 license
• Dockerfile: Yes https://github.com/SillyTavern/SillyTavern/blob/release/Dockerfile
• Demo:

---

• Summary: locally installed user interface that allows you to interact with text generation LLMs, image generation engines, and TTS voice models. Our goal is to empower users with as much utility and control over their LLM prompts as possible, embracing the steep learning curve as part of the fun.

---

• Notes: Multi-modal ai support
  https://docs.sillytavern.app/

---

• Alternative to / Libhunt link: https://www.libhunt.com/r/SillyTavern
• Screenshots: images, brand logo