# model app **URL:** https://discuss.88.io/c/own-interface/desktop-model/313.md [Latest](https://discuss.88.io/latest.md) · [Categories](https://discuss.88.io/categories.md) · [Tags](https://discuss.88.io/tags.md) --- ## [AI Cluster](https://discuss.88.io/t/ai-cluster/1865)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 4\ **Last updated:** [March 20, 2026, 1:48am UTC](https://discuss.88.io/t/ai-cluster/1865 "2026-03-20T01:48:04Z")
AI Teaming Your Own AI Your Own Chat The cheapest chat model on owner compute is qwen3.5:4b-instruct you can compare with chatgpt4.2 qwen3.5:4b-instruct chatgpt5.3 The only AI you can trust is your own AI. by "AI" we mean neural network based artificial intelligence technology. Owner Comput… --- ## [AI for All](https://discuss.88.io/t/ai-for-all/611)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 8\ **Last updated:** [January 15, 2026, 11:15am UTC](https://discuss.88.io/t/ai-for-all/611 "2026-01-15T11:15:34Z")
Power of AI for Everyone There is Nothing wrong with AI, there is Everything wrong with concentration of AI. Current AI Problems Deep Learning based Artificial Intelligence requires massive amount of data and compute power, fitting perfectly into the existing Cloud Computing model, substantially… --- ## [AI Model 2026](https://discuss.88.io/t/ai-model-2026/1989)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [August 23, 2026, 12:08am UTC](https://discuss.88.io/t/ai-model-2026/1989 "2026-08-23T00:08:58Z")
Qwen3-Reranker-4B - 32K context window Example replacing Hindsight defaults: Embeddings: BAAI/bge-small-en-v1.5 Reranking: cross-encoder/ms-marco-MiniLM-L-6-v2 - English Only --- ## [AI Node 2026](https://discuss.88.io/t/ai-node-2026/1988)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 2\ **Last updated:** [August 17, 2026, 1:27am UTC](https://discuss.88.io/t/ai-node-2026/1988 "2026-08-17T01:27:03Z")
2026 Qwen3.6 35B weights distributed across four Nvidia RTX5060ti 16GB using NVFP4 for maximum performance and capacity. Supports uo tp 256K context but designed for up to 32 concurrent conversations each with average 32K context. Average context per conversation Approximate concurrent convers… --- ## [MiniCPM](https://discuss.88.io/t/minicpm/1975)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [July 11, 2026, 12:21am UTC](https://discuss.88.io/t/minicpm/1975 "2026-07-11T00:21:13Z")
MiniCPM5-1B GitHub - OpenBMB/MiniCPM: MiniCPM5-1B: A SOTA 1B on-device LLM, small yet powerful. · GitHub Very smart small 1B model, but to prevent occasion loopsm enforce strict max\_tokens or step ceilings. Agent Compatibility To make the MiniCPM5-1B model compatible with OpenAI's agentic ecosyst… --- ## [Nextcloud MCP](https://discuss.88.io/t/nextcloud-mcp/1973)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [July 8, 2026, 11:15pm UTC](https://discuss.88.io/t/nextcloud-mcp/1973 "2026-07-08T23:15:04Z")
Nextcloud MCP Currently 13 Apps 1. Notes (7 tools) Tool Description nc\_notes\_get\_notes List all notes nc\_notes\_get\_note Get note by ID (with optional body/content) nc\_notes\_create\_note Create a new note nc\_notes\_update\_note Update an existing note nc\_notes\_delete\_note Delete a note … --- ## [LLM Wiki](https://discuss.88.io/t/llm-wiki/1970)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [July 3, 2026, 12:08pm UTC](https://discuss.88.io/t/llm-wiki/1970 "2026-07-03T12:08:12Z")
Desktop GitHub - nashsu/llm\_wiki: LLM Wiki is a cross-platform desktop application that turns your documents into an organized, interlinked knowledge base — automatically. Instead of traditional RAG (retrieve-and-answer from scratch every time), the LLM incrementally builds and maintains a persiste… --- ## [Hermes Agent](https://discuss.88.io/t/hermes-agent/1965)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 3\ **Last updated:** [July 3, 2026, 12:07pm UTC](https://discuss.88.io/t/hermes-agent/1965 "2026-07-03T12:07:04Z")
UI Projects as of 26-06-29 Stars Contributors Update GitHub - fathah/hermes-desktop: Desktop Companion for Hermes Agent · GitHub 12.8K 50 10 hr GitHub - Euraika-Labs/pan-ui: Pan by Euraika — a self-hosted AI workspace for Hermes Agent. Chat, skills, extensions, memory, profiles, and runtim… --- ## [AI Performance](https://discuss.88.io/t/ai-performance/1585)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 2\ **Last updated:** [July 3, 2026, 10:30am UTC](https://discuss.88.io/t/ai-performance/1585 "2026-07-03T10:30:42Z")
Distributed AI workloads can be distributed across many processors. Distribution Distrubuted Llama GitHub - b4rtaz/distributed-llama: Distributed LLM inference. Connect home devices into a powerful cluster to accelerate LLM inference. More devices means faster inference. Exolabs https://exol… --- ## [AI Data Store](https://discuss.88.io/t/ai-data-store/1967)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [July 2, 2026, 8:13am UTC](https://discuss.88.io/t/ai-data-store/1967 "2026-07-02T08:13:51Z")
LLM Wiki RAGFlow OpenViking --- ## [MCP](https://discuss.88.io/t/mcp/1966)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [July 2, 2026, 8:10am UTC](https://discuss.88.io/t/mcp/1966 "2026-07-02T08:10:18Z")
Health --- ## [OpenClaw](https://discuss.88.io/t/openclaw/1947)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 1\ **Last updated:** [May 9, 2026, 2:46am UTC](https://discuss.88.io/t/openclaw/1947 "2026-05-09T02:46:12Z")
OpenClaw Intro OpenClaw Architecture, Explained: How It Works https://sausheong.com/dissecting-openclaw-733213e9c853 OpenClaw Conatiner Docker - OpenClaw Sandboxing - OpenClaw Isolate docker further with LXC or KVM or both. OpenClaw Plugins GitHub - karanuppal/openclaw-credential-vault · GitHu… --- ## [AI Samples](https://discuss.88.io/t/ai-samples/1939)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 2\ **Last updated:** [March 26, 2026, 11:06pm UTC](https://discuss.88.io/t/ai-samples/1939 "2026-03-26T23:06:49Z")
These sample uses the following Alias Teams: insight.aunsw.88.io --- ## [Qwen](https://discuss.88.io/t/qwen/1604)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 6\ **Last updated:** [March 7, 2026, 2:52am UTC](https://discuss.88.io/t/qwen/1604 "2026-03-07T02:52:17Z")
Qwen Recommended Qwen Model qwen3-vl:30b-a3b-instruct-q8\_0 (62GB) - for apps requiring 8-bits Legacy Qwen Models Still supported but not recommended for new installs. qwq qwen2.5 qwen2.5-vl MoE Models Qwen 3 MoE models need more bits in quantisation, try keep to q8 or above for MoE models. Q… --- ## [Duix Avatar](https://discuss.88.io/t/duix-avatar/1882)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 2\ **Last updated:** [February 4, 2026, 1:31am UTC](https://discuss.88.io/t/duix-avatar/1882 "2026-02-04T01:31:39Z")
Team Compute Avatar Video You can join a team to provide avatar video service or you can be a user, either way it is a lot of fun and rewarding. Duix Avatar Multi-language Support: Scripts support eight languages - English, Japanese, Korean, Chinese, French, German, Arabic, and Spanish. --- ## [Kokoro](https://discuss.88.io/t/kokoro/1771)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 1\ **Last updated:** [February 4, 2026, 1:24am UTC](https://discuss.88.io/t/kokoro/1771 "2026-02-04T01:24:46Z")
--- ## [Standard MCP Nodes](https://discuss.88.io/t/standard-mcp-nodes/1889)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 2\ **Last updated:** [February 3, 2026, 3:04am UTC](https://discuss.88.io/t/standard-mcp-nodes/1889 "2026-02-03T03:04:00Z")
Bundled MCP Servers External MCP Server Internal MCP Server --- ## [Weather MCP](https://discuss.88.io/t/weather-mcp/1888)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 1\ **Last updated:** [February 2, 2026, 4:03am UTC](https://discuss.88.io/t/weather-mcp/1888 "2026-02-02T04:03:57Z")
OPen Source Weather --- ## [AI Video](https://discuss.88.io/t/ai-video/1883)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [January 31, 2026, 12:46am UTC](https://discuss.88.io/t/ai-video/1883 "2026-01-31T00:46:18Z")
Video Quality The quality gap between open software AI and commercial software AI is actually not that big ... For example, Compute Owners have access to Duix Avatar which is open software. Below is a sample from Open Software on avatarvideo.aunsw.88.io (which all compute owners have access to)… --- ## [GLM](https://discuss.88.io/t/glm/1881)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 1\ **Last updated:** [January 21, 2026, 4:03am UTC](https://discuss.88.io/t/glm/1881 "2026-01-21T04:03:50Z")
GLM is a family of LLM released by Z.ai --- ## [Reranking Models](https://discuss.88.io/t/reranking-models/1880)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [January 19, 2026, 2:52am UTC](https://discuss.88.io/t/reranking-models/1880 "2026-01-19T02:52:37Z")
Qwen3 Reranker A specialised reranker model can boost search quality. We are using the Qwen3 Reranker as default in 26.07 release. 4B Model Out of the 3 sizes, 0.6B, 4B and 8B we are using the 4B as default. 4B is half the size of 8B but performs only sightly worse than 8B in some cases and is … --- ## [Embedding Models](https://discuss.88.io/t/embedding-models/1581)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 2\ **Last updated:** [January 18, 2026, 4:38am UTC](https://discuss.88.io/t/embedding-models/1581 "2026-01-18T04:38:28Z")
Embedding The word "Embeddings" use in the AI context are numerical representations of real-world objects that machine learning uses to represent complex data and relationships. Unlike traditional "records" in a database that represent just the "value" of the data, embeddings can provide a lot of i… --- ## [Model Context Protocol](https://discuss.88.io/t/model-context-protocol/1879)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [January 17, 2026, 5:44am UTC](https://discuss.88.io/t/model-context-protocol/1879 "2026-01-17T05:44:39Z")
MCP with Team Compute Model Context Protocol (MCP) is a fundamental part of Team Compute, as it allows models to discover what services are available and provides a standard interface to those services: Tools - available actions for the models Resources - available data for the models Prompts - av… --- ## [Machine Translation](https://discuss.88.io/t/machine-translation/1421)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 4\ **Last updated:** [January 2, 2026, 10:16am UTC](https://discuss.88.io/t/machine-translation/1421 "2026-01-02T10:16:22Z")
Translation for All Demo translate.aunsw.88.io LLM Gemma LTEngine uses Google Gemma 3 GitHub - LibreTranslate/LTEngine: Local AI Machine Translation. Powered by LLMs. LibreTranslate compatible. 🚀 LTEngine - LLM powered local machine translation - General - LibreTranslate Community libretranslate/… --- ## [LLM UI](https://discuss.88.io/t/llm-ui/1847)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [November 23, 2025, 6:05am UTC](https://discuss.88.io/t/llm-ui/1847 "2025-11-23T06:05:52Z")
kobold GitHub - LostRuins/koboldcpp: Run GGUF models easily with a KoboldAI UI. One File. Zero Install. --- ## [LLM Proxy](https://discuss.88.io/t/llm-proxy/1846)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [November 23, 2025, 5:08am UTC](https://discuss.88.io/t/llm-proxy/1846 "2025-11-23T05:08:38Z")
ollamaflow https://ollamaflow.com/ GitHub - ollamaflow/ui: UI for OllamaFlow GitHub - ollamaflow/apiexplorer: API explorer for OllamaFlow olla GitHub - thushan/olla: High-performance lightweight proxy and load balancer for LLM infrastructure. Intelligent routing, automatic failover and unif… --- ## [Ollama](https://discuss.88.io/t/ollama/1127)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 10\ **Last updated:** [November 21, 2025, 4:32am UTC](https://discuss.88.io/t/ollama/1127 "2025-11-21T04:32:02Z")
Beginners Model Providers Ollama is easy to use but slow in delivering new features. It supports a lot of models: Other Model Providers for beginners: LocalAI Model Interfaces Open-WebUI GitHub - open-webui/open-webui: User-friendly AI Interface (Supports Ollama, OpenAI API, ...) · GitHub … --- ## [AI Hardware](https://discuss.88.io/t/ai-hardware/1748)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 3\ **Last updated:** [November 20, 2025, 7:43am UTC](https://discuss.88.io/t/ai-hardware/1748 "2025-11-20T07:43:04Z")
If you are buying new AI hardware for sharing with other Compute Owners, as of 2025-06-01 the following AI hardware are recommended due to their ability to deliver the maximum performance from minimum power consumption: Ram Size Hardware Model 16 GB Nvidia RTX 5060 Ti 64 GB Apple M4 Pro … --- ## [Multi-Agent](https://discuss.88.io/t/multi-agent/1818)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 1\ **Last updated:** [October 21, 2025, 11:45pm UTC](https://discuss.88.io/t/multi-agent/1818 "2025-10-21T23:45:25Z")
Team Compute is not just about teams of humans but also teams of agents. --- ## [EgoBlur](https://discuss.88.io/t/egoblur/1752)
**Author:** [@admin](https://discuss.88.io/u/admin)\ **Replies:** 0\ **Last updated:** [June 16, 2025, 1:11pm UTC](https://discuss.88.io/t/egoblur/1752 "2025-06-16T13:11:58Z")
Blur by Default Team Compute attempts to blur images and videos as much as possible for all Compute Rewards. 1. EgoBlur Egoblur is currently being used for both Human Face and Vehicle Number. https://www.projectaria.com/tools/egoblur/ GitHub - facebookresearch/EgoBlur: This repository contains a … [Next page](https://discuss.88.io/c/own-interface/desktop-model/313.md?page=1)