---
title: "raft vs llm-app"
type: "comparison"
canonical_url: "https://www.graphcanon.com/compare/nvidia-raft-vs-pathwaycom-llm-app"
tools: ["nvidia-raft", "pathwaycom-llm-app"]
---

# raft vs llm-app

*GraphCanon updated Jul 11, 2026*

## Verdict

Pick raft when raft is primarily Cuda; llm-app is Jupyter Notebook; pick llm-app when llm-app is primarily Jupyter Notebook; raft is Cuda.

[raft](https://docs.rapids.ai/api/raft/stable/) reports 1.0k GitHub stars, 240 forks, and 448 open issues, last pushed Jul 11, 2026. [llm-app](https://pathway.com/developers/templates/) has 59k stars, 1.4k forks, and 10 open issues, last pushed Jul 5, 2026. Figures are from public GitHub metadata via [raft's repository](https://github.com/NVIDIA/raft) and [llm-app's repository](https://github.com/pathwaycom/llm-app).

| | [raft](/tools/nvidia-raft.md) | [llm-app](/tools/pathwaycom-llm-app.md) |
| --- | --- | --- |
| Tagline | RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing hig | Ready-to-run cloud templates for RAG, AI pipelines, and enterprise search with live data. |
| Stars | 1,026 | 59,068 |
| Forks | 240 | 1,432 |
| Open issues | 448 | 10 |
| Language | Cuda | Jupyter Notebook |
| Adopt for | - | llm-app offers pre-configured cloud deployment templates designed specifically for creating AI-driven applications such as chatbots and machine learning projects leveraging Hugging Face models. It supports direct integrz |
| Persona | - | - |
| Runtime | - | - |
| License | Apache-2.0 | MIT |
| Categories | LLM Frameworks, Vector Databases, Data & Retrieval | LLM Frameworks, Vector Databases, Data & Retrieval |

## Trust and health

_Sourced signals - not a safety guarantee. No winner column._

| | [raft](/tools/nvidia-raft.md) | [llm-app](/tools/pathwaycom-llm-app.md) |
| --- | --- | --- |
| Days since push | 0d | 5d |
| Open issues (now) | 448 | 10 |
| Full report | [trust report](/tools/nvidia-raft/trust.md) | [trust report](/tools/pathwaycom-llm-app/trust.md) |

## Decision facts: llm-app

- **Requirements:** Requires Docker; The tool is Docker-friendly and designed to ensure synchronization with cloud-based storage solutions among others.
- **Adopt for:** llm-app offers pre-configured cloud deployment templates designed specifically for creating AI-driven applications such as chatbots and machine learning projects leveraging Hugging Face models. It supports direct integrz

## Choose when

### Choose raft if…

- raft is primarily Cuda; llm-app is Jupyter Notebook.
- License: raft is Apache-2.0, llm-app is MIT.
- Tags unique to raft: clustering, anns, gpu, building-blocks.

### Choose llm-app if…

- llm-app is primarily Jupyter Notebook; raft is Cuda.
- License: llm-app is MIT, raft is Apache-2.0.
- Requirements: Requires Docker; The tool is Docker-friendly and designed to ensure synchronization with cloud-based storage solutions among others..
- Tags unique to llm-app: vector-database, llm, hugging-face, retrieval-augmented-generation.
- - You need a ready-to-run solution that directly integrates with various data sources like Sharepoint, Google Drive, S3, Kafka, PostgreSQL, and live APIs.

## When NOT to use raft

- LLM Frameworks: Avoid a framework for a single prompt-and-retrieve call; the abstraction can cost more than it saves.
- Vector Databases: Don't reach for a dedicated vector DB under ~100k vectors; pgvector on your existing Postgres is simpler to operate.
- Data & Retrieval: Skip a heavy ingestion framework when your corpus is small and static; a script plus the embedding API is enough.

## When NOT to use llm-app

- - You require custom deployment configurations that extend beyond the pre-set cloud templates available through llm-app.
- - There’s a need for tightly integrated support with data sources or APIs not explicitly mentioned, such as specialized CRM systems (Salesforce), which may lack direct template support in llm-app.

## Common questions

### What is the difference between raft and llm-app?

raft: RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing hig. llm-app: Ready-to-run cloud templates for RAG, AI pipelines, and enterprise search with live data.. See the comparison table for live GitHub stats and shared categories.

### When should I choose raft over llm-app?

Choose raft over llm-app when raft is primarily Cuda; llm-app is Jupyter Notebook; License: raft is Apache-2.0, llm-app is MIT; Tags unique to raft: clustering, anns, gpu, building-blocks.

### When should I choose llm-app over raft?

Choose llm-app over raft when llm-app is primarily Jupyter Notebook; raft is Cuda; License: llm-app is MIT, raft is Apache-2.0; Requirements: Requires Docker; The tool is Docker-friendly and designed to ensure synchronization with cloud-based storage solutions among others.; Tags unique to llm-app: vector-database, llm, hugging-face, retrieval-augmented-generation; - You need a ready-to-run solution that directly integrates with various data sources like Sharepoint, Google Drive, S3, Kafka, PostgreSQL, and live APIs.

### When should I avoid raft?

LLM Frameworks: Avoid a framework for a single prompt-and-retrieve call; the abstraction can cost more than it saves. Vector Databases: Don't reach for a dedicated vector DB under ~100k vectors; pgvector on your existing Postgres is simpler to operate. Data & Retrieval: Skip a heavy ingestion framework when your corpus is small and static; a script plus the embedding API is enough.

### When should I avoid llm-app?

- You require custom deployment configurations that extend beyond the pre-set cloud templates available through llm-app. - There’s a need for tightly integrated support with data sources or APIs not explicitly mentioned, such as specialized CRM systems (Salesforce), which may lack direct template support in llm-app.

### Is raft or llm-app more popular on GitHub?

llm-app has more GitHub stars (59,068 vs 1,026). Stars measure visibility, not whether either tool fits your constraints.

### Are raft and llm-app open source?

Yes - both are open-source projects on GitHub (raft: Apache-2.0, llm-app: MIT).

### Where can I find alternatives to raft or llm-app?

GraphCanon lists graph-backed alternatives at [raft alternatives](/tools/nvidia-raft/alternatives) and [llm-app alternatives](/tools/pathwaycom-llm-app/alternatives) ([raft markdown twin](/tools/nvidia-raft/alternatives.md), [llm-app markdown twin](/tools/pathwaycom-llm-app/alternatives.md)), ranked by typed relationship edges rather than popularity votes.

### Is there a machine-readable version of this comparison?

Yes. The markdown twin at [this comparison](/compare/nvidia-raft-vs-pathwaycom-llm-app.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers.

### Which is better maintained, raft or llm-app?

raft: Very active. llm-app: Very active. Compare maintenance labels, days since push, and release cadence in the trust section below - stars alone do not measure maintenance.

### Where are the full trust reports for raft and llm-app?

GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [raft trust report](/tools/nvidia-raft/trust); [llm-app trust report](/tools/pathwaycom-llm-app/trust).

---

**Machine-readable endpoints**

- JSON: [`/api/graphcanon/graph?tool=nvidia-raft`](/api/graphcanon/graph?tool=nvidia-raft)
- LLM index: [/llms.txt](/llms.txt)
- Full corpus: [/llms-full.txt](/llms-full.txt)

_GraphCanon - The knowledge graph for AI development. https://www.graphcanon.com/_