Start with the
ecological question.
Specify the species, study area and research purpose. Review a proposed analysis plan with explicit assumptions.
Output · A documented study planAI-assisted ecological research
Understand where
species can thrive.
And why.
We’re building a research platform that brings climate, habitat and species observations into one reproducible modelling workflow.
Researcher-led. Evidence-traceable. Starting with butterflies.
FROM OBSERVATIONS TO EVIDENCEPlanned workflow
Species observationsGBIF + researcher field records
Climate & habitatEnvironmental context, aligned in space and time
SCIENTIFIC QUALITY GATE
Question the data.
Then build the model.
Comparable modelsClimate · Habitat · Combined
Traceable inputs. Documented decisions. Reproducible outputs.
Species distribution modellingScientific quality assessmentOpen research infrastructure
The research challenge
A model is only as useful
as the evidence behind it.
Scattered datasets, uneven sampling and undocumented choices make ecological modelling difficult to assess and repeat. EcoLeptr Lab is being designed to make those decisions visible, from the first record to the final map.
Designed for
Academics and PhD researchers
Universities and conservation organisations
Consultancies and independent labs
Working example · GBIF data audit
One butterfly.
A visible first output.
A small, reproducible audit of public occurrence records shows the kind of quality questions EcoLeptr Lab is designed to surface.
Papilio machaon · Türkiye
GBIF snapshot · 9 October 2026Occurrence data audit
300records examined
37without a stated coordinate uncertainty
245flagged as coordinate rounded
262from one dataset in this sample
First 300 API results out of 950 reported matches. These signals call for review; they are not automatic exclusions.
Inspect the complete example reportThis is a fixed demonstration sample, not a species distribution model or an assessment of conservation status. It does not represent all records for the species. Read the report and method.
The planned platform
One question.
A complete research workflow.
From a research question to an analysis another researcher can inspect, understand and repeat.
Start with the
Bring the evidence
into alignment.
Combine occurrence records with climate, land cover and field data. Check taxonomy, coordinates, dates and sampling patterns.
Output · A traceable data auditCompare models.
Test their limits.
Evaluate established methods using spatial validation, a simple baseline and sensitivity checks suited to the data.
Output · An assessed model comparisonMake every result
inspectable.
Export suitability and uncertainty maps alongside methods, source references, settings and runnable analysis code.
Output · A reproducible research packageClimate + habitat
Three perspectives.
A clearer ecological picture.
The first modelling scope compares what each source of environmental information contributes.
Climate
Explore climatic suitability using temperature, precipitation and relevant seasonal variables.
Where are climate conditions suitable?
Present conditions + future scenariosHabitat
Assess environmental context through land cover, terrain and species-relevant habitat information.
Where could suitable habitat occur?
Matched to the study’s date and scaleCombined
Bring climate and habitat together, then test whether the added information improves generalisation.
What changes when we consider both?
Evaluated against the simpler modelsFuture projections will identify their climate scenarios and habitat assumptions. Habitat suitability describes environmental conditions; population trends require additional evidence.
Scientific quality, built in
See what your
study is missing.
Every proposed quality check is designed to connect an observed issue to its possible effect and a practical next step.
When the evidence is insufficient, the system should explain what additional data are needed before modelling can proceed.
Read our scientific commitmentsILLUSTRATIVE QUALITY REVIEWExample only
01Sampling concentrated near roads
- Evidence
- Records cluster around accessible locations.
- Potential effect
- The model may learn observer access as well as ecological preference.
- Next step
- Review sampling effort, background selection and spatial validation.
02Observation and habitat dates differ
- Evidence
- Older occurrences are paired with a recent land-cover layer.
- Potential effect
- Environmental values may not describe the habitat when the species was recorded.
- Next step
- Align time periods or report the mismatch and test sensitivity to filtering.
03Host-plant information is missing
- Evidence
- No suitable host-plant layer is available for a host-dependent species.
- Potential effect
- Environmental suitability may overstate where the species could establish.
- Next step
- Document the limitation and prioritise host-plant data collection or expert review.
Our scientific commitments
Researcher judgment.
Transparent computation.
Automation should make research decisions easier to examine and results easier to reproduce.
Researchers set the direction
Researchers review study boundaries, assumptions and analysis choices. Recommendations come with reasons and limitations.
Calculations remain traceable
Versioned analytical tools produce numerical results. Data snapshots, settings and software versions accompany each analysis.
Validation goes beyond one score
Spatial evaluation, sensitivity checks and uncertainty reporting help assess where a model can be trusted.
Sharing follows evidence and consent
Open research outputs retain provenance and attribution. Unpublished records and sensitive species locations need appropriate access controls.
Planned Claude integration
An assistant across the research workflow.
We plan to use Claude to structure research questions, coordinate approved analysis tools and explain verified outputs. Numerical models will run in reproducible analytical environments, with key scientific choices available for researcher review.
INITIAL RESEARCH FOCUS / LEPIDOPTERAFocused by design
Butterflies first.
Biodiversity ahead.
Our starting point is butterfly research, with a pilot scope shaped around expert access, available field observations and environmental data.
Where evidence allows, species-specific considerations will include host plants, flight periods and dispersal. Expansion to other taxa will follow validation of each group’s data and modelling requirements.
First milestoneA small, expert-reviewed butterfly pilot.
The evidence foundation
Open data.
Accountable use.
Planned data sources and reporting references. Each analysis should identify exactly what was used, when it was accessed and how it was processed.
OCCURRENCE RECORDS
GBIF
Species observations with source attribution, download references and quality checks.
Visit source CLIMATE SCENARIOSWorldClim / CMIP6
Climate layers and future scenarios, with model, period and resolution documented.
Visit source LAND COVERESA WorldCover
Land-cover information assessed for its date, spatial scale and ecological relevance.
Visit source REPORTING FRAMEWORKODMAP
A reference for structured documentation of data, models, assessment and prediction.
Visit referenceThese are intended sources and references, not partner endorsements. Data availability, licences and suitability will be assessed for each study.
A shared record of ecological knowledge.
Our longer-term goal is an open registry of permitted datasets, documented corrections and expert-reviewed models. Every contribution should retain its source, version and uncertainty, with publication controlled by its contributor.
Development roadmap
Build carefully.
Validate before expanding.
EcoLeptr Lab has an early data-audit demonstration. The modelling service is not yet publicly available.
- 01CURRENT FOCUS
Research design
Define the first user workflows, scientific quality requirements and butterfly pilot scope.
- 02PLANNED
Validated pilot
Build the data audit and modelling workflow; compare results with expert analyses and independent observations.
- 03PLANNED
Open research registry
Introduce permission-aware sharing, model versioning and updates assessed against held-out evaluation data.
- 04LONG-TERM DIRECTION
Conservation evidence
Expand to additional taxa and assemble evidence packages to support expert conservation assessments.
Progress measured through research quality.
Time saved, issues detected, reproducibility and independent validation will guide development.
Who we build for
Research users.
Institutional buyers.
EcoLeptr Lab is exploring a paid research workflow for individuals and organisations that need repeatable biodiversity analyses.
Individual researchers
Academics and PhD students who need transparent, time-saving analysis support.
Research and conservation teams
Universities, conservation organisations and independent labs coordinating studies and evidence reviews.
Applied ecology teams
Consultancies and environmental assessment projects that need documented methods and auditable outputs.
The intended revenue model is access to the research platform and team workflows. Scope, packaging and pricing will be shaped with pilot users; no paid service is offered yet.
Help shape the first pilot
Bring a research question.
We’ll explore the fit.
We are inviting researchers and organisations to discuss an early butterfly-focused pilot. Tell us what you study and where the current workflow slows you down.
No pilot partnership, response time or product access is promised by submitting this form.
Scope & transparency
A few important
distinctions.
Can I run a species distribution model today?
The public modelling service is in development. This site describes the product scope and scientific approach; the workflow and quality-review examples are illustrative.
How will models improve over time?
The planned progression starts with comparisons between established methods and parameter tuning. Later updates will be versioned and evaluated on held-out data before replacing an earlier model. Expert feedback and new observations will inform improvements.
Will EcoLeptr Lab assign official conservation categories?
The long-term aim is to prepare traceable evidence for expert assessment. Modelled habitat suitability alone is insufficient to establish population decline or an official IUCN Red List category.
Will all research data become public?
The planned open registry will include only records and outputs that can be shared under their licences and contributor permissions. Unpublished research and sensitive locations require separate access choices.
Our purpose