SurveyX¶
external · status: active · focus: literature · discipline: general · started: 2025
Project page: https://github.com/IAAR-Shanghai/SurveyX
Source: projects/landscape/surveyx.yml
Positioning¶
An academic survey-automation system (arXiv:2502.14776) that generates domain-specific surveys from a paper title plus retrieval keywords. Open-source repository ships only the offline processing path; full-version capabilities (real-time online search, multimodal parsing) are gated behind the hosted surveyx.cn service.
Distinctive contribution¶
Two-tier deployment — an open offline pipeline plus a commercial online service — making the field's tension between research reproducibility and product gating explicit. The offline release is usable as a methodological baseline.
Evaluation scores¶
| Dimension | Score (0–3) | Note |
|---|---|---|
| Lifecycle coverage | 1 | Three stages around literature synthesis and drafting. |
| Autonomy level | 3 | Runs from title + keywords to survey; offline version is autonomous given the reference set. |
| Architectural transparency | 2 | arXiv:2502.14776 documents the method; offline code is public; online crawler + paper-DB system are closed. |
| Inputs supported | 1 | Open version: user-uploaded Markdown references only. Full feature set requires the hosted service. |
| Outputs / reproducibility | 1 | Open path is reproducible given user-supplied references; hosted-service runs are not. |
| Internal evaluation | 2 | ArXiv paper reports comparative evaluation; the hosted-service variant has not been independently audited. |
| Openness | 1 | No declared license in repository metadata; substantial functionality reserved for the commercial service. |
| Maturity / traction | 2 | 970 stars; active development; live hosted service at surveyx.cn. |
| Cross-family policy | 0 | No cross-family policy in open release. |
| Runtime assurance | 1 | Offline path has reference-grounding only; full crawler + paper-DB closed. |
| Cross-platform portability | 1 | Offline (markdown) + hosted-service (closed) paths. |
Scored on 2026-05-18. See the evaluation rubric.
Tags¶
Pipeline stages: literature-discovery literature-synthesis paper-drafting
Architectural features: multi-agent rag-knowledge-base iterative-loop
Inputs: survey-title keywords reference-markdown
Outputs: long-form-survey citations
Data sources: user-provided-markdown
Knowledge sources: user-provided
Limitations¶
- Open-source release deliberately omits the most capability-relevant components (crawler, paper DB, multimodal parser).
- No declared open-source license.
- Output quality on the open path depends entirely on the references the user supplies.
Related projects in this catalog¶
Papers describing this project¶
- SurveyX: Academic Survey Automation via Large Language Models — Liang, X., Yang, J., Wang, Y., Tang, C., Zheng, Z., Song, S., et al. (2025). arXiv. arXiv:2502.14776
Related references (literature catalog)¶
- Wu, J. et al. (2025). Agentic Reasoning: A Streamlined Framework for Enhancing LLM Reasoning with Agentic Tools
wu2025agenticreasoning