ASReview is an open source screening tool for the systematic literature review process. The core mechanism is to combine active learning with manual annotation iteration to shorten the title and abstract screening cycle while ensuring the recall rate.

ASReview

Core parameters and statistics

Specific technical parameters (such as model size, context length, supported file formats, input and output restrictions, etc.) are subject to the official product page. It is recommended that users verify the latest technical specifications and system requirements before choosing to ensure that they match their own usage scenarios.

User and market recognition

ASReview's recognition mainly comes from academic and evidence-based research scenarios, rather than the mass consumer market.

Research Community Recognition: Its GitHub community has been active for a long time, and its stars and forks maintain a medium to high position among open source research tools, indicating that it has been continuously used in the real review process.

Organization Adoption Characteristics: The official website emphasizes "Trusted by the top-notch organizations" and provides access to scientific publishing and learning materials, reflecting its stable demand in high-evidence-density industries such as scientific research, medical care, and policy evaluation.

Reproducible value: Compared with the purely manual screening process, ASReview's traceable annotation and iterative sorting mechanism are more suitable for teams that require audit link and method transparency.

Cost advantage

C-side/individual: ASReview is provided in open source form, and individual researchers can start it with zero license cost; the main investment is the learning curve and local context configuration time.

API/Developer: The code warehouse and documentation are open. Developers can do secondary development according to the research process. The explicit cost is low, but they need to bear the responsibility of model strategy selection and data pipeline maintenance.

Enterprise/Institution: The public page does not provide a unified enterprise price list. If organizational-level support, compliance integration or customized deployment is required, the fees and service boundaries are subject to official real-time communication results.

Hidden costs: The true cost of a systematic review project is usually determined by annotation quality, team collaboration efficiency, and review rework rate. The value of ASReview lies in reducing the amount of invalid screening, rather than simply lowering software subscription fees.

Main functions

• Active learning sorting: Dynamically adjust the sorting of candidate documents based on labeled samples, put highly relevant documents in the front, and prioritize the release of manual time.
• Interactive screening process: Support researchers to provide feedback while screening, and continuously iterate model judgments.
• Performance comparison and method comparison: Different models/strategies can be compared to help the team choose a more robust screening path.
• Crowdscreen collaboration mechanism: allows expert groups to participate in screening, improving consistency and reducing single-person bias.
• Open source scalability: adapts to the reproducible needs of research teams, and can be combined with own data and process specifications for secondary development.

Model and version evolution

ASReview's version evolution focuses on "stabilizing the screening experience + continuously repairing key failure points in the research process."

Mainline release

• v3.0.7 (2026-06-03): Fixed the ranking table loss problem and updated the front-end dependencies.
• v3.0.6 (2026-06-01): Fix database lock problem and improve session stability.
• v3.0.5 (2026-03-30): 3.0 branch maintenance version, continuing the mainline functions and stability optimization.

Version relationship

The 3.0.x series reflects continuous maintenance within the same product main line, rather than a sudden change in the direction of capabilities. For the production research process, what is more important is version stability and consistency of replication experiments, rather than the pursuit of frequent switching to new branches.

Technical advantages

Mechanism: Use active learning for literature screening, and take "human-computer collaborative annotation" as the core.

Effect: In the high-cardinality candidate document pool, highly relevant records can be pushed to the front more quickly, thereby reducing the first round of manual browsing and shortening the screening cycle.

Applicable scenarios: When the team faces thousands to tens of thousands of candidate records, the benefits of this sorting and prioritization mechanism are most obvious; when the candidate set is small or the problem is unclearly defined, the benefits will decrease.

Engineering Advantages: Apache-2.0's open source license and public code structure lower the threshold for method auditing, making it easier for scientific research teams to conduct replication experiments, strategy comparisons, and process traces.

How to use

Entrance	Typical steps	Applicable teams
Official site and documents	Prepare literature data -> Create screening project -> Manually label the first batch of samples -> Iterative screening and export results	Scientific research and evidence assessment team
Open source warehouse deployment	Clone warehouse -> Install dependencies -> Start local/server instance -> Configure filtering strategy	Research engineering team
Collaborative screening process	Set inclusion and exclusion criteria -> Assign multiple people to review -> Summarize conflicts and review	Institutional-level systematic review project

When implementing, it is recommended to first use historical projects to do a small sample playback to calibrate the inclusion standards and review rules, and then expand to formal projects to reduce the amplification effect of early strategic deviations on the results.

Product Pricing

The pricing model is subject to the official real-time page. Usually a freemium or subscription system is used, and basic functions can be used for free. Advanced functions or high-frequency use require paid subscriptions, and users are advised to evaluate the optimal solution based on actual usage.

Application scenarios

• Medical and Public Health System Review: Prioritize high-relevant studies in large-scale literature libraries to reduce the investment in human and day screening.
• Social Science Evidence Integration: Used for policy evaluation and evidence mapping to improve the consistency of cross-topic retrieval.
• Corporate Research and Intelligence Analysis: Use active learning for preliminary screening of patents, technical reports and industry literature to shorten the pre-research preparation cycle.

The common premise for these scenarios is that the team has clearly defined inclusion/exclusion criteria and has the most basic data cleaning and review process in place.

Applicable people

• Scientific researchers and systematic review authors: It is necessary to improve screening efficiency in the interpretable process.
• Research Engineer and Methodology Team: An open source solution that is scalable, reproducible, and re-development is needed.
• Institutional Level Evidence Assessment Team: Requires multi-person collaboration, audit trail, and process standardization.

The boundary of incompatibility is equally clear: if the task is not systematic evidence screening, the sample size is very small, or there are no clear screening criteria, it is difficult to give full play to the process advantages of ASReview.

Summary and Outlook

ASReview's core competitiveness lies in transforming "highly repetitive screening labor" in systematic review scenarios into "an iterable human-computer collaboration process" and improving the transparency of the research process through open source and reproducible mechanisms. It has obvious value in academic and evidence research scenarios, but it also places higher demands on the methodological discipline and process governance of the team.

The current uncertainties mainly lie in the boundaries of commercialization support and the disclosure of institutional-level service terms. In practice, you can first use historical projects as a pilot to compare the differences in recall rate, time-consuming, and review consistency between manual screening and the ASReview process; after reaching expectations, expand to cross-research group or institution-level deployment, and verify version stability, data governance responsibilities, and collaboration specifications before expansion.

Version Info

• ASReview v3.0.7 ：GitHub Releases The latest stable version, including ranking list loss repair and front-end dependency updates. （2026-06-03）
• ASReview v3.0.6 ：Fixed issues related to database locks and improved the stability of filtering sessions. （2026-06-01）
• ASReview v3.0.5 ：The 3.0 branch is a continuous maintenance version used to consolidate the main line of filtering and data flow functions. （2026-03-30）