Academic Commercialization Assessment Agent
A multi-agent LLM pipeline; a research topic can be submitted or a paper PDF attached.
Verified
Core facts
Read by a script from the official API, not rewritten by a model.
- License
- MIT
- Languages
- Python
- Maintainer
- shuxiachai
- Verified
What it is
The project is a multi-agent LLM pipeline. The project is built with Python, CrewAI, FastAPI and a build-free JavaScript client. The project was created by shuxiachai. The project is licensed under MIT.
Main features
Users can submit a research topic or attach a paper PDF, and choose the report language.
Deterministic code retrieves and validates sources first.
The weighted total is computed by code, not the model.
Users can follow progress, inspect citations and reliability warnings, and export Markdown.
An interrupted run can be resumed from its last validated checkpoint.
Paid requests get durable receipts.
Getting started
Python 3.11 or 3.12 is used for the CI-tested environment.
The default test suite does not call providers.
In the quick start, dependency installation needs network access.
Copy the example environment file to a local environment file.
Unused placeholder keys should be removed from the configuration.
Real analysis incurs provider usage.
PDF extraction is a paid operation.
The test suite and browser smokes make no paid requests.
git clone https://github.com/shuxiachai/academic-commercialization-agent.git
cd academic-commercialization-agent
uv syncLLM_PROVIDER=qwen
DASHSCOPE_API_KEY=your-key
QWEN_MODEL=qwen3.8-27b
QWEN_API_BASE=https://dashscope.aliyuncs.com/compatible-mode/v1
TAVILY_API_KEY=your-search-keyuv run uvicorn api.main:app --reload
# Browser: http://localhost:8000 (sample reports: http://localhost:8000/samples)
# Alternative CLI (starts real provider work):
uv run academic_agent --topic "solid-state batteries for electric vehicles"Who it is for
A report supports research triage. A valid citation ID does not establish that the source supports the claim. In the architecture, three evidence specialists run in parallel.
In the architecture, a stage is not a promise of exactly one model request. A resumed run is a new linked run, and the original is never overwritten. In Sources, a saved source can be found by ID or keyword and its saved text read.
The frozen baseline contains 10 topics × 3 live repetitions. The frozen baseline's results are different checks, not a combined accuracy score. The uncited-numeric proxy does not measure all hallucinations.
Tool execution or valid JSON does not prove that a claim is supported. Supplementary retrieval remains zero-call shadow mode in production. Run URLs carry 128 bits of randomness and act as read capabilities.
Do not publish private run URLs. Changing a code-owned run additionally requires that run's owner/admin code. An account's keys stay in its browser and are never stored by the server.
Deployments use one application replica / one Uvicorn worker. Reported costs cover Crew nodes, and do not cover helper, PDF or search charges.
Related tools
Related tools listed on this site.
FeelFish
An AI writing workspace for long-form fiction that keeps characters, worldbuilding, chapters, and context in one project.
Loomi
Loomi’s current entry is an AI creative workspace describing image, video and audio creation from a sentence or a photo.
MiniMax
Multimodal foundation-model company whose site introduces text, video, speech and music models plus products such as MiniMax Code, Design, Audio and Talkie, and an API entry.
QwenPaw
Personal AI assistant and agent workstation running in your own environment. Its primary description presents local/cloud deployment, multiple channels and extension through Skills and MCP.
ZCode
AI coding tool from Zhipu. Its site describes combining GLM models, coding agents and an existing toolchain for planning, coding, review and multi-step development tasks.
知乎直答
Zhihu Zhida’s official public description presents an AI information service centered on Zhihu community content with additional data sources, describing a multi-agent system and source tracing.
纳米AI
Nano AI is an AI search and agent assistant whose site says it can generate videos, reports and slides from one sentence and integrates several large models.
Devin Desktop
The site describes Devin Desktop as the home for every agent you run, a command center to manage agents locally or in the cloud; its FAQ states it is the new name for Windsurf, built on Windsurf's IDE foundation.
Originality.AI
Content-review platform whose site describes AI-text detection, plagiarism checking, writing quality and fact-checking entry points.
Sources
This page is written from the official sources below. Open them to check the original text.
- GitHub READMEhttps://github.com/shuxiachai/academic-commercialization-agentFetched
Verified ·
This page is written by AI from the official sources above and checked automatically. This site has not tested it. Found a mistake? Tell us.