Browser-extension AI writing assistant that drafts emails and messages in your style, fixes grammar and chats with the page you are on.
CAJAL
Open-source local scientific paper generator that drafts academic papers offline and has AI judges review them.
What is CAJAL?
CAJAL is an open-source project for writing scientific papers on your own computer instead of through a hosted chatbot. The GitHub page describes a Qwen 27B model fine-tuned for academic writing that can run offline on a single RTX 3090 graphics card, and the project connects to any Ollama-compatible local model. It drafts papers in the usual academic layout, looks up references through arXiv and CrossRef, and passes each draft to a panel of AI judges it calls a tribunal, which gives a quick peer-review style score. The repository also mentions a CognitionBoard and a token-compression feature, plus kits for agent frameworks and editors such as VS Code and Jupyter. Researchers and graduate students use it to get a first draft or a rough review without sending unpublished work to a cloud service. The code is free to use, so the real cost is the hardware and the time spent checking the output.
Use cases
- Draft a structured research paper
- Pull references from arXiv and CrossRef
- Score a draft with AI reviewers
- Write academic text offline
- Add paper writing to agent workflows
Who it's for
Why use CAJAL
- Keeps prompts and unpublished research on your own machine rather than a third-party API.
- No subscription or per-token bill, since the model runs on hardware you control.
- The judge panel gives a first-pass critique before you ask a colleague to read the draft.
- Framework and editor kits let you call paper generation from tools you already use.
CAJAL features
- Runs offline against a local Ollama-compatible model
- Fine-tuned Qwen 27B model aimed at academic writing
- Standard academic paper structure from abstract to references
- Reference lookup through arXiv and CrossRef
- Tribunal of several AI judges that scores each paper
- 6x4 CognitionBoard and 2.7x token compression
- Integration kits for LangChain, CrewAI, AutoGen and LlamaIndex
- Editor kits for VS Code and Jupyter
CAJAL integrations
How to get started with CAJAL
- Install a local model runtime such as Ollama and download a compatible model.
- Clone the CAJAL repository from GitHub and follow its setup steps.
- Give it a research topic and let it draft the paper section by section.
- Read the judge scores and the references it found, then edit and verify the draft yourself.
CAJAL limitations
- Self-hosted: you supply the GPU and the local model
- The large fine-tuned model targets a high-end consumer GPU such as an RTX 3090
- Generated papers need careful human checking of claims and citations
CAJAL pricing
CAJAL is open source, so you can use and self-host it for free.
Prices can change and may vary by country or exclude tax. Check the official pricing page.
CAJAL alternatives
The main CAJAL alternatives listed on Debuly are Elicit, Consensus, AgentDock, HyperWrite, Jenni and Yomu AI. The table compares pricing and platforms; order is A–Z.
| Tool | Pricing | Starting price | Platforms |
|---|---|---|---|
| CAJAL | Open source | — | Self-hosted |
| AgentDock | Freemium | $12/mo | Web, Chrome extension, API, CLI |
| Consensus | Freemium | $12/mo | Web, API, MCP |
| Elicit | Freemium | $49/mo | Web, API |
| HyperWrite | Freemium | $16/mo | Web, Chrome extension |
| Jenni | Freemium | $12/mo | Web, Chrome extension |
| Yomu AI | Freemium | $10/mo | Web |

AgentDock
AI platform whose Dock suite pairs a citation-aware writing editor, a Chrome extension and a prompt library, with a customer-engagement AI employee alongside.
FreemiumVisit AgentDock (official site)
Consensus
AI search engine for research that answers questions from peer-reviewed papers, with every claim linked back to its source.
FreemiumVisit Consensus (official site)
Elicit
AI research assistant that searches 138 million papers, extracts data into tables and runs systematic reviews and cited reports.
FreemiumVisit Elicit (official site)
HyperWrite
AI writing assistant with sentence autocomplete, web research and citations, email replies, and a library of single-purpose writing tools.
FreemiumVisit HyperWrite (official site)Jenni
AI academic writing workspace with autocomplete, chat, a source library and citations in one editor for papers and essays.
FreemiumVisit Jenni (official site)
Yomu AI
Yomu AI is an academic writing workspace that drafts with citations, edits papers, chats with PDFs and runs a plagiarism check.
FreemiumVisit Yomu AI (official site)
CAJAL FAQ
How much does CAJAL cost?
The repository lists no price. It is open-source software, so you pay only for your own hardware and electricity.
Is there a free plan or a paid plan?
There are no plans. The project is free to download and run, and it has no tiers or subscriptions listed.
Does CAJAL need an internet connection?
The GitHub page says it can run offline on an RTX 3090. Reference lookup through arXiv and CrossRef would need a connection.
Can I trust the papers CAJAL writes?
Treat every output as a draft. Check the claims, data and citations yourself before sharing or submitting anything.
What are alternatives to CAJAL?
Hosted tools such as Paperpal, Jenni AI, SciSpace and Elicit cover similar ground. CAJAL differs by running locally on an open model.
Sources & verification
Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.
| Source | Type | Accessed |
|---|---|---|
| github.com/Agnuxo1/CAJAL | official |
Image credits
- Logo: none collected yet, so a monogram is shown.
- Cover: none collected yet, so a generated tile is shown.
Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.


