For communities
Your territory. More possibilities.
Our commitment: zero-cost prefeasibility for communities worldwide. Register free while we develop the service.
CAPYBARA CODE RULES.
The AI-powered platform
that gets natural capital
investment done.
From community potential to credible projects—connecting people, spatial intelligence, evidence and investment, and making project development accessible and affordable for communities.
Starting in Peru. Built for communities everywhere.
For the Fronterra team
One place to move projects forward.
Bring project work, spatial intelligence, evidence and review into one shared operating environment.
01
Why this exists
The first step should be within reach.
Communities hold knowledge of their territories. Exploring a carbon project also requires technical expertise, time and resources that can be difficult to access.
OpenCarbon brings these pieces into one workflow: the community's information, the evidence behind an assessment and the people who review it. Our ambition is to make initial prefeasibility available to communities at zero cost, starting in Peru.
What is available now
Free registration, document submission and a qualitative first screen based on the information provided. A submission enters the team's review process. It does not yet receive an automated spatial prefeasibility study.
02
The pipeline in code
Boundary in.
Evidence for a decision out.
From a territory boundary to two independent lines of enquiry: restoration under ARR, and avoided deforestation under REDD+. Here is how the intended workflow fits together.
Illustrative workflow
boundary.geojson
01 / Input
{
"type": "Feature",
"geometry": {
"type": "Polygon",
"coordinates": [[
[-74.21, -8.54], [-74.19, -8.54],
[-74.19, -8.52], [-74.21, -8.52],
[-74.21, -8.54]
]]
},
"properties": {
"example_only": true,
"source": "synthetic_boundary"
}
}spatial_screening.py
02 / Evidence
# Pseudocode for a future processing recipe.
# Sources, dates and permissions need confirmation.
boundary = validate_geometry(submitted_boundary)
sources = select_sources(
geography=boundary,
dates="to be confirmed",
permitted_use="screening"
)
evidence = screen_land_history(boundary, sources)
evidence.record_sources_and_assumptions()
evidence.flag_missing_information()review_pathways.py
03 / Methodology
# Pseudocode. Assess both pathways independently.
arr = review_pathway(
framework="VM0047", evidence=evidence
)
redd_plus = review_pathway(
framework="VM0048 / applicable modules",
evidence=evidence
)
redd_plus.require("applicable allocation / activity data")
for pathway in [arr, redd_plus]:
pathway.methodology_check = "pending specialist review"
pathway.withhold_carbon_and_revenue_estimates()screening_record.json
04 / Review record
{
"example_only": true,
"ARR": {
"methodology_check": "pending",
"carbon_tco2e": null,
"revenue": null
},
"REDD+": {
"methodology_check": "pending",
"carbon_tco2e": null,
"revenue": null
},
"dataset_versions": [],
"reviewer": null,
"status": "evidence_required"
}The target is traceable evidence for specialist review. A real assessment must identify its boundary version, sources, dates, method and reviewer before a result can be released.
03
The prefeasibility path
A clear path.
A person at the decision.
This is the workflow we are building toward. Its first step is open today; spatial checks and report production still require development and specialist work.
Today’s first screen
Defined rules interpret the information supplied in the form. This is qualitative screening, not a satellite analysis, a language-model assessment or a carbon calculation.
Register the territory
AvailableShare the community's information, a territory map or boundary file, photographs and the required supporting documents.
Prepare and check the boundary
In developmentA usable polygon can support spatial analysis after validation. A scanned or photographed map needs assisted digitisation first. Automated overlap checks are being built on the spatial database foundation.
Run spatial screening
Planned integrationConnect repeatable analyses of land cover, forest change and other relevant layers. Preserve the source dates and assumptions, and flag missing information for each pathway.
Review the evidence
Specialist workA specialist reviews applicability and limitations. Quantitative outputs require suitable data, a defensible method and technical approval. A map or an automated score cannot provide that approval.
Respond to the community
Target reportThe intended report explains ARR and REDD+ potential separately, supporting evidence, uncertainties and next steps. Today, communities receive an acknowledgement while the team reviews the request and develops the service.
A request may lead to a conversation about a Fronterra project. Registration alone is not project approval, a financing decision or a promise of carbon credits.
04
What makes an assessment useful
Evidence before estimates.
These principles guide the product and the team's review. The complete automated workflow is still being built.
No number without a basis.
Credit-volume and revenue estimates need more than a screening signal or a reviewer's name. They require appropriate evidence, methodology-specific calculations, uncertainty treatment and review. Missing inputs mean the figures stay withheld.
Projections are not allocations.
For avoided unplanned deforestation under VM0048/VMD0055, the applicable jurisdictional allocation and project activity data matter. A locally modelled deforestation projection does not replace the required allocation process or data.
VT0007 defines the allocation procedure. Clark Labs and TerraCarbon developed UDef-ARP to support its application; it is a separate, GPL-licensed project, not an OpenCarbon integration or an allocation issued to a project.
The right data for the right use.
Review the geography, period, version, licence and permitted use of each source. An older dataset may still be valid historical evidence. A public map may support screening without being sufficient for project registration.
Keep the evidence with the decision.
The intended assessment record connects a boundary version, data sources, methods, assumptions and the technical review. Updates create a new version, preserving the basis of the previous decision.
Prefeasibility is an early step.
Full Fronterra feasibility also requires the technical work, project-appropriate FPIC and community engagement, and IFC Performance Standards due diligence. Screening is not certification, validation or an investment decision.
05
Under the surface
The tools, and where they stand.
OpenCarbon is the team's working environment. Its infrastructure holds records and files; specialist tools support the analyses we plan to connect.
- Static application on VercelLive
- HTML, CSS and JavaScript power two separately deployed portals: a public community portal and Fronterra OS, the internal workspace with its own sign-in.
- Supabase and PostgreSQLLive
- Accounts, application records and access rules support submissions and team work. Community-uploaded documents use private object storage. Some controlled project documents remain in external repositories referenced by the team.
- Qualitative screening rulesLive
- The current first screen applies defined rules to declared information. It does not call a language model to calculate carbon potential or process satellite imagery.
- PostGIS and the spatial workspaceFoundation deployed
- The geographic extension and spatial workspace tables are deployed. Automated overlap analysis and geographical screening are being built on that foundation.
- Google Earth EngineSoon
- A specialist platform for satellite imagery and repeatable regional analyses. Work in this tool remains outside OpenCarbon; automated Earth Engine jobs are not connected to community submissions.
- TerrSet and specialist GIS toolsExternal workflow
- The team can download original submission files and upload analyst outputs in OpenCarbon. Land-change modelling and detailed spatial work remain under specialist control in external tools; this file-based handoff does not run those tools automatically.
- MINAM and Verra reference dataReview required
- GeoBosques provides official Peruvian forest-monitoring information. Verra provides jurisdictional data for its allocation process. The exact product, period and permitted use must be checked for each assessment; possession of a dataset alone does not establish applicability.
Connecting a tool means more than calling its API. The processing recipe, credentials, cost controls, dataset provenance and review path need to work together before results can be released.
06
The commitment
Cost to register
for communities
Start in Peru.
Build for communities everywhere.
Registration is free today. Our goal is also to make community prefeasibility available at zero cost, while Fronterra manages the underlying expertise and processing costs.
The pilot starts in Peru. The team confirms participation, scope and next steps with each community. Communities in other regions can register their interest for review; registration does not guarantee a study.
We aim for initial screening in minutes when inputs are suitable. That is a development target. We will publish turnaround times once the complete workflow has been tested and measured.