Rastrum
About Rastrum
Rastrum is an open-source biodiversity observation platform that combines computer vision, audio analysis, and expert curation to identify species from photos, videos, audio, and indirect evidence.
Making biodiversity knowledge accessible to everyone — from field researchers to curious hikers — so every observation can drive conservation.
Our Vision
The Problem
Species identification tools are fragmented across platforms, locked behind language barriers, and rarely work offline. Most existing models are trained on data from North America and Europe, leaving megadiverse regions like Latin America critically underserved.
The Solution
Rastrum unifies photo, audio, and video identification into a single open platform powered by multiple AI models and validated by community experts. Every observation feeds a growing dataset that strengthens regional accuracy and fuels real conservation action.
Who is behind Rastrum
A small, honest team. We tell you who we are because you're trusting us with your location, your photos, and your time.
Artemio Padilla
Solo dev and maintainer
Builds and maintains Rastrum on nights and weekends. Based in Mexico. Reachable via GitHub Issues or Discussions.
@ArtemioPadillaPhotos are optional and not required to contribute.
Want your name here? We accept code, translations, taxonomic review, field testing, and bug reports. How to contribute.
How it's funded
We believe you should know exactly who pays for what you use.
Current status
Rastrum is currently unfunded. There is no company, no investors, and no advertising. Hosting and infrastructure costs are covered by the operator out of pocket. Volunteer time is the largest input.
What it costs to run
- Supabase Free tier (database, auth, storage) — currently $0/mo
- Cloudflare R2 (media + ML model hosting) — single-digit dollars per month
- Domain (rastrum.org) — about $12 per year
- AI identification calls — pay-as-you-go via your own BYO API keys, or sponsored slots in a community pool
AI identification (Claude Vision, PlantNet) uses your own API keys by default. Your photos are sent to those providers under your account; we never store the keys server-side. See the Privacy page for the full data flow.
When the project starts taking outside money, this section becomes a public ledger. We commit to publishing every recurring cost and every donation source as line items. Funding.
Governance and licensing
Decisions are made in the open, by the people who do the work.
License
Code is MIT-licensed. The server (deployed Edge Functions) is AGPL-3.0 — if you run a public Rastrum instance with modifications, you must share those modifications. Each observation carries its own Creative Commons license, chosen by the contributor (CC BY, CC BY-NC, or CC0).
LICENSEHow decisions get made
Roadmap items live as GitHub Issues with public labels. Anyone can open an issue, comment, or send a pull request. Architectural changes go through a written spec under docs/specs/modules/ before code lands. The maintainer has the final call when no consensus emerges, but the bias is always toward shipping the most-debated path.
GitHub IssuesHow to contribute
Code, translations, taxonomic review, field testing, bug reports, and small donations of compute via the platform sponsorship pool are all welcome and equally valued.
By the numbers today
Live counts from public observations. Updated each time the site is built.
—
Public observations
—
Public observer profiles
12
Distinct species recorded
Last updated: May 29, 2026
How It Works
Observe
Capture photos, audio, or video of any species in the field.
Identify
Our AI pipeline analyzes your evidence using PlantNet, BirdNET, and Claude Vision.
Contribute
Validated observations join a growing open dataset for conservation and science.
Roadmap
Foundation
- Astro skeleton + Supabase schema
- Photo ID MVP with PlantNet
Intelligence
- Claude Vision integration
- Observation log + GPS tagging
Community
- Audio ID with BirdNET
- Expert validation system
Full Platform
- Video analysis support
- Offline mode + Darwin Core export
Next Frontier
- Regional ML models from community data
- AR species overlay
Conservation Impact
Built With
Astro
Fast, modern web framework for static-first delivery
Supabase
Open-source backend for database, auth, and storage
PlantNet
Specialized AI for plant and fungi identification
BirdNET
Neural network for bird and wildlife audio recognition
Claude Vision
Advanced multimodal AI for general species identification
Press and mentions
No press coverage yet. If you write about Rastrum, we'd love to hear about it — open an issue and we'll add the link here.
Open Source
Rastrum is AGPL-3.0-licensed and community-driven. Every observation, every validation, every line of code strengthens the global biodiversity knowledge commons.
GitHub