CorpusKit is an open source research portal. Search across reports, papers and other source material, ask questions of the collection and trace connections from one source to another. Answers cite the material they draw on, down to the relevant passage.
See how it worksOpen source under the Apache 2.0 licence
Source code, setup guides and issue tracker on GitHub
A portal is built around one organisation’s research collection. People can ask questions across it, search for specific material, browse the Library and inspect relationships in the knowledge map. Manage is where administrators add and maintain the material.
AskEnter a question in Ask. CorpusKit searches the collection and returns an answer with citations, sources and a confidence rating. Deep research is available when a broad question needs to be broken into smaller ones.
Question entered in Ask · whole collection
Seagrass recovered more slowly at sites where winter turbidity remained above the long-term median.
At those sites, light at the canopy fell below the minimum required by Zostera muelleri on 63 per cent of monitored days.
Seagrass recovery was slower where winter turbidity remained above the long-term median. A related report attributed an estimated two thirds of the fine sediment reaching the estuary after the floods to bank erosion in the upper catchment.
…sites where mean winter turbidity remained above the long-term median showed significantly slower recovery…
…bank erosion in the upper catchment contributed an estimated two thirds of the fine sediment load reaching the estuary…
Estuarine seagrass condition assessment 2019 to 2023
The report covers five years of condition monitoring across five estuaries, including seagrass extent, density and light availability.
Catchment sediment sources after the 2022 floods
The report traces fine sediment in the estuary to its catchment sources and ranks the available management options.
Turbidity and light limitation in flood-affected estuaries
The report links high winter turbidity after floods to less light at the seagrass canopy and slower recovery.
Recreational water quality guidelines review
The review examines recreational water quality thresholds and the monitoring evidence behind them.
CorpusKit links claims in an answer to the source passages that support them. Quality scores show how relevant the retrieved material is, how closely the answer follows it and whether enough context was found.
Seagrass recovered more slowly at sites where winter turbidity remained above the long-term median. At those sites, light at the canopy fell below the minimum required by Zostera muelleri on 63 per cent of monitored days.2
Sites where mean winter turbidity remained above the long-term median showed significantly slower recovery of seagrass extent over the two years following the floods. At these sites, light at the canopy fell below the minimum requirement for Zostera muelleri on 63 per cent of monitored days.
Matched passage · page 41CorpusKit’s application code is published under the Apache 2.0 licence. Clone the repository, connect your research collection and run the portal on infrastructure you manage.
The repository includes instructions for local and container-based setup, along with the tests, build tasks and deployment configuration used by the project.
Content storage, retrieval and generation require an account with a connected knowledge service. The Apache 2.0 licence covers the CorpusKit application. Your infrastructure, connected services and research material have their own terms.
Changes are submitted through pull requests. Before a pull request can merge, it must pass typechecking, linting, formatting, tests and a full build. These checks apply to everyone.