RAG Frameworks · raphaelmansuy

edgequake

EdgeQuake is a Rust-based knowledge graph RAG framework that transforms documents into intelligent graph structures for retrieval-augmented generation. It combines vector search with graph traversal to handle complex multi-hop reasoning and relationship queries more effectively than traditional RAG systems.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/raphaelmansuy/edgequake

GitHub stars

234

Forks

Rust

Primary language

Apache-2.0

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	raphaelmansuy/edgequake
Owner	raphaelmansuy
Primary language	Rust
License	Apache-2.0 — OSI-approved
Stars	2k
Forks	234
Open issues	26
Latest release	v0.15.0 (2026-07-07)
Last updated	2026-07-07
Source	https://github.com/raphaelmansuy/edgequake

What edgequake is

EdgeQuake implements the LightRAG algorithm in Rust, decomposing documents into entity-relationship knowledge graphs stored in PostgreSQL (pgvector + Apache AGE). It provides six query modes (naive, local, global, hybrid, mix, bypass) with REST API, multi-tenant isolation, and support for multiple LLM providers (OpenAI, Anthropic, Gemini, Ollama, etc.).

Quickstart

Get the edgequake source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/raphaelmansuy/edgequake.git2cd edgequake3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Enterprise Document Intelligence & Multi-Hop Reasoning

Organizations needing to answer complex, cross-document questions (e.g., regulatory compliance, contract analysis, competitive intelligence). The knowledge graph captures relationships that vector-only RAG misses.

High-Concurrency, Low-Latency Document APIs

SaaS platforms requiring sub-500ms query latency and 1000+ concurrent users. EdgeQuake's Rust backend and hybrid query modes are optimized for production scale without traditional RAG's memory bloat.

Domain-Specific Knowledge Management

Healthcare, legal, manufacturing, and research domains where entities, relationships, and thematic clustering matter. Custom entity types (up to 50 per workspace) and community detection support specialized taxonomies.

Implementation considerations

PostgreSQL 16+ with pgvector and Apache AGE extensions required; no embedded database option. Plan for database migration/provisioning upfront.
LLM provider integration (OpenAI, Ollama, etc.) needed for entity extraction and query answering; budget for LLM token costs or self-hosted model overhead.
Multi-pass gleaning (second-pass entity extraction) increases latency and LLM calls; tune extraction depth vs. latency tradeoff based on document complexity.
Vision mode (PDFs as images) via GPT-4o/Claude adds cost and latency (~2-3s per page); use selectively for complex tables or fallback intelligently to text mode.
Custom entity types and domain glossaries require upfront domain modeling; the system ships with 5 presets (General, Manufacturing, Healthcare, Legal, Research).

When to avoid it — and what to weigh

Simple Keyword Lookup Use Cases — If your workload is purely keyword-based retrieval (e.g., FAQ search), traditional vector RAG or full-text search is simpler and lower operational overhead.
Minimal Infrastructure / Stateless Preference — EdgeQuake requires PostgreSQL 16+ (pgvector + Apache AGE extensions). If you want zero-dependency, cloud-native vectorstore-only solutions, this adds complexity.
Closed-Source or Proprietary Requirement — EdgeQuake is Apache 2.0 licensed, openly developed. If your organization restricts open-source dependencies or requires vendor SLA guarantees, commercial RAG platforms may be mandated.
Unproven Graph Extraction Quality for Your Domain — Entity and relationship extraction rely on LLM accuracy. If your documents are highly specialized (e.g., niche scientific notation, industry jargon), entity extraction may hallucinate or miss domain-specific concepts.

License & commercial use

Apache License 2.0 (Apache-2.0): permissive OSI-approved license. Allows commercial use, modification, and distribution with attribution and indemnification clauses. No patent grant limitations noted in the data provided.

Apache 2.0 permits commercial use without explicit vendor approval. However, ensure your organization's legal review covers patent indemnity and liability clauses. EdgeQuake itself is community-developed; no commercial SLA or vendor support model noted in the data. Use at own risk or contact maintainers for enterprise support arrangements.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Strong
License clarity	Clear
Deployment complexity	Moderate
DEV.co fit	Strong
Assessment confidence	High

Security considerations

Multi-tenant workspace isolation enforced at query/deletion layers and authentication/authorization built-in, but extent and default security posture are not detailed in the data. Verify: default credentials, encryption at rest/in transit, audit logging completeness, and rate limiter configuration. No public security audit or vulnerability disclosure policy mentioned. Evaluate alongside your threat model before production use.

Alternatives to consider

LangChain + LlamaIndex

Popular Python RAG frameworks with broader ecosystem; lower barrier to entry for non-systems engineers. No built-in knowledge graph; traditional vector RAG. More integrations but slower for high-concurrency workloads.

Native graph database with established enterprise presence. Requires managing Neo4j infrastructure separately; not as tightly integrated as EdgeQuake's PostgreSQL + AGE approach. Better for existing Neo4j deployments.

Microsoft Copilot Studio / Semantic Kernel

Enterprise-backed, closed-source solutions with SLA guarantees. Managed service reduces operational burden. Vendor lock-in and licensing costs; less transparency and customization.

Software development agency

Build on edgequake with DEV.co software developers

EdgeQuake is open-source, production-ready, and actively maintained. Start with Docker in 5 minutes or discuss a custom deployment with our team.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

MiniRAG LightRAG RAG-Anything

Related on DEV.co

Explore the category and the services that help you build with it.

RAG Frameworks AI Application Development API Development Cloud Deployment Talk to DEV.co

edgequake FAQ

Can I run EdgeQuake without PostgreSQL?

No. The data provided shows pgvector + Apache AGE as required storage backends. Edge-only or fully embedded options are not mentioned.

What LLM providers are supported?

OpenAI, Anthropic, Gemini, Mistral, Ollama, LM Studio, xAI, Azure, and Vertex AI are listed. Specific version compatibility and feature parity across providers not detailed in the data.

How does hybrid query mode differ from traditional RAG?

Hybrid combines vector similarity search with graph traversal of entities and relationships. Traditional RAG relies on vectors alone. EdgeQuake claims 5x faster latency and 3x more entity extraction, but benchmark methodology and dataset not provided.

Is there a managed / hosted version?

Not mentioned in the data. The project is self-hosted via Docker or built from source. Contact maintainers to inquire about commercial hosting options.

Work with a software development agency

Adopting edgequake is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate rag frameworks software in production.

Ready to Build Intelligent Document Systems?

EdgeQuake is open-source, production-ready, and actively maintained. Start with Docker in 5 minutes or discuss a custom deployment with our team.

Talk to DEV.co Schedule a Consultation