MyCaptionAI

Open Source (Llama / Mistral)

19Tools
14Categories
19Active

Model Family

3 tools

Llama

Open-weight or provider model strategy

Meta's open model family and tooling foundation for a large share of the open-weight ecosystem.

Build on Meta's open foundation modelsFine-tune for domain-specific tasksRun locally or on any cloud
ViewVisit ↗

Mistral

Open-weight or provider model strategy

Mistral AI's open-weight and commercial model family used across local, hosted, and enterprise deployments.

Deploy efficient multilingual modelsUse Mixtral for MoE performanceOn-prem enterprise deployment
ViewVisit ↗

Gemma

Open-weight or provider model strategy

Google's lightweight open model family for local, cloud, and deployable AI applications.

Run Google's lightweight open modelsDeploy on edge and mobile devicesFine-tune with limited resources
ViewVisit ↗

Assistant

1 tool

Le Chat

Daily Open Source (Llama / Mistral) workflows

Mistral's assistant workspace for chat, search, document analysis, Canvas, code interpreter, and custom agents.

Conversational AI in French and EnglishAgentic web search and canvasDaily productivity assistant
ViewVisit ↗

Developer Platform

1 tool

Mistral AI Studio

Production app setup and governance

Mistral's developer console for API keys, playground testing, agents, fine-tuning, evaluation, and usage monitoring.

Fine-tune Mistral models on your dataExperiment with prompt engineeringDeploy custom models to production
ViewVisit ↗

Coding Agent

1 tool

Mistral Vibe

Developer automation

Mistral's terminal-native coding agent product in the open model coding stack.

Generate web apps from promptsRapid UI prototyping with AIPreview and iterate on code live
ViewVisit ↗

Local Runtime

3 tools

Ollama

Local or private inference

Popular local and cloud model runner for pulling, running, serving, and integrating open models.

Run Llama and Mistral models locallyServe private API endpointsQuick model experimentation
ViewVisit ↗

LM Studio

Local or private inference

Desktop and developer runtime for chatting with, loading, serving, and connecting local models privately.

Chat with models offline on desktopBenchmark different model sizesServe local OpenAI-compatible APIs
ViewVisit ↗

llama.cpp

Local or private inference

Core C/C++ runtime for quantized local inference, GGUF models, and OpenAI-compatible local serving.

Run quantized models on CPUEmbed inference in C++ appsMinimal-dependency local inference
ViewVisit ↗

Model Hub

1 tool

Hugging Face Hub

Model discovery and evaluation

Central hub for discovering, testing, hosting, and deploying open models and datasets.

Discover and compare open modelsHost datasets and model cardsTry models with instant demos
ViewVisit ↗

Model Hosting

1 tool

Replicate

Hosted open-model inference

Cloud platform for running, fine-tuning, and deploying open models via APIs.

Run open models via simple API callsScale GPU inference without DevOpsFine-tune models in the cloud
ViewVisit ↗

Inference Server

1 tool

vLLM

Local or private inference

High-throughput open-source serving engine for production open model inference.

High-throughput production servingContinuous batching for GPUsDeploy OpenAI-compatible endpoints
ViewVisit ↗

Self-Hosted UI

1 tool

Open WebUI

Self-hosted user experience

Self-hosted AI interface for local and cloud models, commonly paired with Ollama and OpenAI-compatible APIs.

Give teams a private ChatGPT-like UIConnect to Ollama and cloud APIsManage users and conversation history
ViewVisit ↗

Local Assistant

1 tool

Jan

Self-hosted user experience

Open-source local-first assistant and model platform with desktop apps, MCP, Jan Hub, and a local API server.

Offline-first AI desktop assistantSwitch between local and cloud modelsBuild RAG pipelines locally
ViewVisit ↗

Local API Stack

1 tool

LocalAI

Local or private inference

MIT-licensed local AI stack and OpenAI-compatible API for self-hosted language, image, audio, and agent workloads.

Self-host multiple model typesDrop-in OpenAI API replacementRun on consumer hardware
ViewVisit ↗

Model Provider

2 tools

DeepSeek

Open-weight or provider model strategy

DeepSeek model ecosystem for efficient reasoning, coding, and open model experimentation.

Efficient reasoning at lower costOpen-weight coding assistanceMath and science problem-solving
ViewVisit ↗

Qwen

Open-weight or provider model strategy

Alibaba's Qwen model family and assistant ecosystem for open and hosted reasoning, coding, and multimodal work.

Multilingual Asian language tasksLong-context document processingCode generation and math reasoning
ViewVisit ↗

Open-Weight Models

1 tool

gpt-oss

Open-weight or provider model strategy

OpenAI's open-weight reasoning models also belong in the wider self-hosted open model stack.

Self-host reasoning models on-premFine-tune on proprietary datasetsRun air-gapped inference
ViewVisit ↗

Inference Provider

1 tool

Groq Cloud

Hosted open-model inference

Low-latency hosted inference option frequently used for supported open models.

Ultra-fast inference for open modelsSub-second response latencyCost-effective high-throughput serving
ViewVisit ↗