RAG Showcase¶
Six modern RAG approaches, compared side by side — each served as an OpenAI-compatible endpoint on a fully-local Atlas stack. Ask one question in Open WebUI and watch the approaches answer in parallel, with a uniform answer, retrieved-context, and metrics surface. It doubles as a reproducible evaluation harness that measures which approach wins on which kind of question.
1. The Six Approaches¶
| Endpoint | Approach | Designed to shine on |
|---|---|---|
vanilla-rag |
Dense top-k retrieval, then a single generation call (the control) | Simple factoids; the baseline |
hybrid-rag |
Weaviate hybrid retrieval (BM25 + dense), then TEI reranking | Exact keyword and identifier queries |
contextual-rag |
Anthropic Contextual Retrieval over context-prefixed chunks | Context-starved chunks |
graph-rag |
LightRAG over extracted entities, relationships, and vector context | Graph-shaped relationship questions |
agentic-rag |
ReAct loop over vector and graph retrieval tools | Multi-hop and comparative questions |
n8n-adaptive-rag |
Low-code workflow that routes by query complexity | Mixed simple-and-complex batches |
The last column is the design intent behind each demo query family, not a measured result — the committed runs contradict some intended contrasts (see the per-query winners).
Any approach can also expose tuned flavors — for example hybrid-rag-high-recall
or graph-rag-fast — that route to the same base approach with reproducible parameter
overrides and their own selectable model alias. See Flavor Tuning.
2. Headline Result¶
The 2026-07-03 ladder ran fourteen approach and flavor aliases across three datasets of increasing structure. The winner shifts with input complexity, which is the point:
| Dataset | Winning configuration | Judge score |
|---|---|---|
| Baseline curated | vanilla-rag-wide |
4.42 |
| Graph-native dossiers | hybrid-rag-high-recall |
4.25 |
| Cyber-threat graph (MITRE ATT&CK) | contextual-rag-high-recall |
3.58 |
Full per-query winners, judge methodology, and raw snapshots are in Evaluation and Results.
3. Documentation¶
-
Get Started
Prerequisites, one-command bring-up, and driving the comparison in Open WebUI.
-
The Approaches
Step-by-step internals, dependencies, and tuning knobs for each of the six.
-
Evaluation and Results
Judge-panel methodology, the dataset complexity ladder, and the measured rankings.
-
Architecture
The plugin seam, LiteLLM, retrieval stores, and workflow services.
4. Fully Local by Default¶
Everything runs on your own machine: local models through Atlas's Ollama provider
(qwen3.6:latest for chat, nomic-embed-text for embeddings, and mistral-small3.2:24b
for LightRAG's graph roles), Weaviate and LightRAG for retrieval, a TEI reranker, and a
local judge panel. No cloud calls are required to run the showcase or reproduce its
results. See the Hardware Sizing guide for minimum and recommended profiles.
The project is also a deliberate test-drive of Atlas as reusable infrastructure. The Atlas Reuse Assessment records what reused cleanly, the seams that were added, and the pinned dependency contracts each integration was verified against.