RagLeap
No CUDA, no 2GB torch. 23MB reranker. 6 vector DBs.
Start new thread

RagLeap - No CUDA, no 2GB torch. 23MB reranker. 6 vector DBs.

by•
The open-source core of RagLeap — a self-hosted RAG chat engine (document ingestion, retrieval, citations) powering RagLeap's full AI business manager: Voice AI, WhatsApp/Telegram/Discord bots, database & CRM integrations, workflow automation, and a private executive assistant. Foundation layer — hosted platform at ragleap.com - antonyrag/ragleap-core

Add a comment

Replies

Best
Maker
📌
I built this because every RAG library I tried assumed I had a GPU. RagLeap's reranker runs on ONNX Runtime — a 23MB quantized model instead of the 2GB+ torch/CUDA stack most RAG frameworks pull in even for CPU-only use. It supports 6 vector backends (pgvector, FAISS, Pinecone, Weaviate, Qdrant, Milvus), 8 embedding providers, and 12+ generation providers with automatic fallback if one goes down. Real per-call cost tracking with monthly budgets. 238 tests, real CI on every PR. Everything in the docs is either live-verified against real infrastructure or explicitly labeled unverified — no marketing claims, no numbers we haven't actually measured. pip install ragleap-rag Contributions welcome — the roadmap, and what we've deliberately said no to, are both documented in the open: github.com/antonyrag/ragleap-core Would love feedback, especially from anyone who's fought GPU dependency bloat before.
Maker

Quickstart for anyone who wants to try it:

pip install ragleap-rag

from ragleap import RagLeap, ProviderConfig, EmbeddingConfig

rag = RagLeap(

database_url="postgresql://user:pass@localhost/mydb",

embedder=EmbeddingConfig(provider="gemini", api_key="YOUR_KEY"),

primary=ProviderConfig(provider="gemini", api_key="YOUR_KEY"),

)

rag.init_schema()

rag.ingest("doc.pdf", open("doc.pdf","rb").read())

answer = rag.ask("your question")

print(answer["answer"])

Works on CPU with 23MB ONNX. No GPU needed.

I also added WhatsApp/Telegram/Discord examples in /examples folder on GitHub.

Which one should I demo next?

Maker

🚀 2 Hour Update

We just hit:

- 4 GitHub stars

- 1 fork

- 141 commits on ragleap-core

pip install ragleap-rag is live if you want to try the CPU-first RAG engine.

Biggest question I'm getting: "How is this different from LangChain?"

Answer: No abstraction layers. 23MB ONNX reranker instead of 2GB torch. Built-in WhatsApp/Telegram/Discord. You own the DB.

What feature should I build next?

A) More vector DBs

B) Better WhatsApp bot templates

C) One-click deploy script

Comment below 👇 I'll build the top voted one this week.

Maker

🗺️ 2026 Roadmap - The 8 Package Vision

Thanks for the support today! Here's where RagLeap is headed:

v0.11.2 DONE: ragleap-rag - CPU-first RAG engine

Q3 2026 NEXT:

- ragleap-graph: Neo4j knowledge-graph RAG

- ragleap-vectorstores: Qdrant, Weaviate, Chroma support

- ragleap-tools: search, code exec, calculators

- ragleap-integrations: MCP-native + WhatsApp/Telegram/Discord

Q4 2026 AFTER:

- ragleap-agents: Role-based crews + tool-calling + HITL

- ragleap-flows: Low-code orchestration

NEW: ragleap-observability: Tracing + hallucination detection + LLM-as-judge evals

ONGOING: ragleap-ops: Docker/K8s/Helm templates

LATER: ragleap-studio: Visual builder UI

All MIT licensed. All CPU-first.

Which package do you want first? 👇

Maker

🚀 24 Hour Update from the maker

Thanks to everyone who tried RagLeap today!

Quick stats:

- GitHub: 4+ stars, 143 commits

- pip install: `pip install ragleap-rag` is live

- Runs on: CPU only. Tested on 8GB RAM laptop

Most common question: "Is this really faster than LangChain?"

Answer: Yes. No 2GB torch download. 23MB ONNX loads in 2 seconds.

I just pushed:

1. Neo4j support → ragleap-graph is WIP

2. 3 new vector DBs → Qdrant, Weaviate, Chroma coming

3. Full roadmap in pinned comment above

If you get it running, drop your use-case below 👇

I'll personally help you debug in comments.

Last 12 hours to upvote. Appreciate the support!

Maker

⏰ Last 3 hours to support RagLeap on PH

Building this solo for 4 months. 143 commits.

Goal: Make RAG accessible on CPU for every dev.

If you believe in "no GPU needed AI", please upvote 🙏

Roadmap is pinned above. `ragleap-graph` with Neo4j starts next week.

Thanks to everyone who commented and starred. This keeps me going.