Practical overview
When it fits
Use Minima when you want to configure and operate document RAG on your own computer or infrastructure, and need to select where the language model runs or how users reach document search.
Who it is for
- People or organizations wanting conversational search over a defined local or on-premises document collection.
- Operators who want to choose between local Ollama inference and a separately deployed OpenAI-compatible LLM.
- Users who want local-document retrieval accessible through a ChatGPT custom GPT or Claude integration.
Documented uses
- Index a configured folder and ask questions grounded in its supported documents.
- Run the language model, reranker, and embedding model locally through Ollama.
- Connect local indexing to a separately deployed OpenAI-compatible LLM.
- Expose local-document search through ChatGPT custom GPTs or Claude using MCP.
Requirements & limits
The repository identifies Minima as open source under the MPL-2.0 license. The documented deployment uses Docker Compose.
- The four documented modes differ in model placement: fully local Ollama, a custom OpenAI-compatible LLM, ChatGPT, or Claude through MCP.
- The indexer can run locally while a custom LLM runs on a server, in the cloud, or on a local machine.
- Supported indexed file types listed in the README are PDF, XLS, DOCX, TXT, Markdown, and CSV.
- Embedding configuration includes both a Sentence Transformer model ID and its dimension, which must match the model.
- ChatGPT integration has USER_ID and PASSWORD environment settings; those settings are distinct from the fully local and custom-LLM configurations.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Configure Minima for fully local Ollama use
Prepare the documented environment settings and Docker Compose command for the isolated local mode.
Before you start
- A cloned Minima project with its .env.sample file
- Docker Compose
- A document folder to index
- An Ollama model and Sentence Transformer embedding model
- A reranker model for Ollama mode
Steps
Create a project-root .env file by copying the variables from .env.sample.
Set LOCAL_FILES_PATH, EMBEDDING_MODEL_ID, and EMBEDDING_SIZE to match the document folder and embedding model.
Set OLLAMA_MODEL and RERANKER_MODEL for the Ollama mode.
Run docker compose -f docker-compose-ollama.yml --env-file .env up --build.
Expected outcome
The documented container setup is configured for fully local RAG, with the language model, reranker, and embeddings running through Ollama.
Guide limitations
- The embedding dimension must match the selected model.
- This guide covers the documented configuration and launch command, not a verified deployment.
Configure Minima with a custom OpenAI-compatible LLM
Prepare the environment settings and Docker Compose command for connecting Minima to a separately deployed LLM.
Before you start
- A cloned Minima project with its .env.sample file
- Docker Compose
- A document folder to index
- A deployed OpenAI-compatible LLM endpoint
Steps
Create a project-root .env file by copying the variables from .env.sample.
Set LOCAL_FILES_PATH, EMBEDDING_MODEL_ID, and EMBEDDING_SIZE.
Set LLM_BASE_URL and LLM_MODEL; add LLM_API_KEY if the endpoint requires authentication.
Run docker compose -f docker-compose-custom-llm.yml --env-file .env up --build.
Expected outcome
The documented custom-LLM container setup connects local indexing to the configured OpenAI-compatible model endpoint.
Guide limitations
- The endpoint must be OpenAI-compatible.
- Ollama and reranker settings are not required for this mode.
- This guide covers the documented configuration and launch command, not a verified deployment.
Compare options for Retrieval-augmented applications
These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.
| Project | Documented purpose | Platform | Access and requirements |
|---|---|---|---|
| Minima | Index a configured folder and ask questions grounded in its supported documents. | Cloud, Docker, Self-hosted | The repository identifies Minima as open source under the MPL-2.0 license. The documented deployment uses Docker Compose. |
| Rag Powered AI Assistant Transforming Data Retrieval And Analysis Across The Web And Pdfs | Retrieve relevant passages from web and PDF content to answer a question. | Web | |
| LangChain in Supabase Edge Functions | Initialize and start a Supabase project locally. | Cloud, Server |
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 12 October 2026
- Original website title
- Minima-AI-Inc/minima · Official homepage title · retained in its source language