llm-app vs manticoresearch
Two of the top projects in RAG and search, side by side: who leads on each score part and who each one suits, then setup, license, activity and what each review found.
llm-app
YAML-configured RAG server that keeps a vector database synced with a document folder
manticoresearch
C++ search database with full-text, vector and hybrid search over SQL and HTTP
Verdict
llm-app ranks higher, 77 to 75. Of the 4 score parts we measure for both, llm-app leads on 1, manticoresearch on 1 and 2 are even.
- Adoptionllm-app leads by 24 points, 98 to 74: a wide gap.
- FreshnessEven at 100.
- Maintenancemanticoresearch leads by 20 points, 78 to 58: a wide gap.
- Agent-readyEven at 0.
Pick llm-app if you are
Teams wanting live, cited Q&A over a document folder without writing code
- Edits, additions and deletions propagate to the vector store as real upserts and deletes
- 8 backends: DuckDB (default, embedded), pgvector, Qdrant, Milvus, ChromaDB, Weaviate, Pinecone, MongoDB Atlas
Pick manticoresearch if you are
Engineers replacing Elasticsearch or adding hybrid search to self-hosted apps
- Hybrid full-text and KNN retrieval in one query, with auto-embeddings and chunking
- Speaks MySQL protocol and HTTP JSON; official clients for nine languages
| What we compare | llm-app | manticoresearch |
|---|---|---|
| Score parts, out of 100 | ||
| Adoption | 98, widely used | 74, popular |
| Freshness | 100, active | 100, active |
| Maintenance | 58, fair | 78, fair |
| Agent-ready | 0, none | 0, none |
| Facts from GitHub and the README | ||
| Stars | 58.7k | 12.1k |
| License | MIT (permissive) | GPL-3.0 (copyleft) |
| Last commit | Nov 2026 | Nov 2026 |
| Last release | Nov 2026 | Sep 2026 |
| Language | Python | C++ |
| Docker | No | Yes |
| GPU | Optional | Not stated |
| arm64 or Apple Silicon | Not stated | Mentioned |
llm-app
serviette watches a folder or remote source (Google Drive, S3/MinIO, SharePoint, FTP/SFTP/WebDAV), parses and embeds documents with Pathway, and writes them to one of 8 vector databases. A separate stateless server hosts a web chat UI on / and a REST API under /api/v1 on port 8989, answering with cited sources. Edits and deletions in the source become upserts and deletes in the store within seconds.
Who it is for: Teams wanting live, cited Q&A over a document folder without writing code
Strengths
- Edits, additions and deletions propagate to the vector store as real upserts and deletes
- 8 backends: DuckDB (default, embedded), pgvector, Qdrant, Milvus, ChromaDB, Weaviate, Pinecone, MongoDB Atlas
- Indexer and server are separate processes sharing only the database; server scales horizontally
- Published benchmarks: 50 GB of text indexed at 14.65 GiB peak PSS, plus FRAMES retrieval results
Weaknesses
- Requires a Pathway license key, free but obtained via GitHub or LinkedIn sign-in
- No conversation memory or user accounts; answers one question at a time
- Default DuckDB backend is single-writer, so it suits laptops and demos only
- Local embeddings pull in PyTorch (about 0.8 GB for the CPU build)
- Python, OpenAI, LiteLLM, SentenceTransformers, Gemini
- Needs Pathway license key, Python with pip 23.1+, OpenAI API key (optional, for generated answers), Vector database (embedded DuckDB by default)
manticoresearch
Manticore Search is a C++ search database with SQL (MySQL protocol) and HTTP JSON interfaces, positioned as an Elasticsearch alternative. It combines full-text search, KNN vector search, hybrid queries and auto-embeddings with chunking, and adds conversational search through SQL CALL CHAT. Storage is row-wise or columnar, with real-time inserts, sharded tables and Galera-based multi-master replication.
Who it is for: Engineers replacing Elasticsearch or adding hybrid search to self-hosted apps
Strengths
- Hybrid full-text and KNN retrieval in one query, with auto-embeddings and chunking
- Speaks MySQL protocol and HTTP JSON; official clients for nine languages
- Columnar storage for datasets larger than RAM; about 40MB RSS when empty
- Replication, sharded tables, built-in auth, backup to S3-compatible storage
Weaknesses
- GPL-3.0 licence may constrain embedding in proprietary products
- Not fully ACID; only isolated transactions and binary logging
- Elasticsearch compatibility covers writes only, not the full query API
- Benchmark claims come from the project's own suite and are workload-specific
- C++, auto-embeddings, LLM-backed CALL CHAT
- Needs Manticore Columnar Library (columnar storage), LLM provider (conversational search)
More in RAG and search
Building with these?
We help teams pick an open-source AI stack and run it in production. Client work never changes a rank.