mcp-inicio-ia
MCP server indexing Markdown docs in SQLite FTS5 with notes and Ollama assistant.
Tested on Linux
Use it
Linux x86_64
You need first
- Runtimes:
python>=3.12 - System packages:
SQLite with FTS5
clone repository to /tmp/mcp-inicio-ia
git clone https://github.com/em4nu3i69dll/mcp-inicio-ia.git /tmp/mcp-inicio-iainstall requirements and package with pip
cd /tmp/mcp-inicio-ia && python3 -m pip install -r requirements.txt && python3 -m pip install --no-deps .
Check that it works
cd /tmp/mcp-inicio-ia && python3 -m mcp_inicio --helpThe output should contain “usage: mcp-inicio”.
First thing to try
cd /tmp/mcp-inicio-ia && python3 -m mcp_inicio prepararFrom the project folder, run python3 -m mcp_inicio preparar to index the documents first.
What it does
- Search local documentsTested
How to use it
- Run
python3 -m mcp_inicio prepararin the project folder to index the documents. - Run
python3 -c "from mcp_inicio.almacen import Almacen; import json; print(json.dumps(Almacen('.').buscar('<your text>'), ensure_ascii=False, indent=2))"to get matching fragments.
- Run
- Save and look up your notesTested
How to use it
- Write your note text into a file.
- Run
python3 -m mcp_inicio nota --titulo "<title>" --archivo <file>to save it. - Run
python3 -c "from mcp_inicio.almacen import Almacen; import json; print(json.dumps(Almacen('.').notas(), ensure_ascii=False, indent=2))"to list saved notes.
- Read an indexed document by its IDTested
How to use it
- Run
python3 -m mcp_inicio prepararto index the documents. - Run `python3 -c "from mcp_inicio.almacen import Almacen; import json; a=Almacen('.'); docs=a.listar(); did=docs['documentos'][0]['id']; print('DOC_ID:'+did); print(json.dumps(a.leer(did), ensure_ascii
- Run
- Ask a local assistant through OllamaNot tested
How far pickfit got
Linux x86_64
- Installs
- Starts
- 3 of 4 features work
- Works offline
macOS: the project says it runs there; pickfit hasn’t tried it.
Good for
- search indexed documents by relevance
- read indexed document by ID with pagination
- save user note from a text file
- expose server context via MCP resource
- prepare research query from a question
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 30 Oct 2026
- Version 1.0.0