Open source
Connectors, automation toolkits, and MCP servers that integrate leading sustainability software with AI agents and modern engineering workflows.
AvailablePython toolkit for automating openLCA calculations, batch runs, parameter updates, and result extraction over the IPC protocol.
Replaces manual desktop work with scripted, repeatable workflows — so a study can be re-run, versioned, and audited rather than rebuilt by hand.
pip install openlca-ipcIn developmentMCP server that connects AI agents to openLCA workspaces, models, calculations, and results through a structured interface.
Exposes LCA operations as typed, inspectable tools so automation stays legible — every agent action maps to a documented model operation.
In developmentMCP server for Brightway2, enabling AI-assisted LCA workflows over databases, methods, and calculations.
Brings the Brightway ecosystem — a widely used programmatic LCA framework — into agent and assistant workflows without hiding the underlying calls.
In developmentMCP server for BioSTEAM models, simulation runs, and scenario analysis with AI-assisted interpretation.
Lets assistants build, run, and interrogate process models while the flowsheet, assumptions, and results stay in plain, reviewable form.
Why it matters
Sustainability analysis is only as credible as its assumptions. Open tooling keeps those assumptions inspectable.
Built on open protocols and interoperable interfaces that avoid vendor lock-in.
Clean APIs, clear documentation, and modular design for fast integration and extension.
Open source by design, community maintained, and transparently governed.
Connects leading sustainability tools and AI agents across ecosystems and workflows.
Architecture
Our infrastructure sits between AI-driven workflows and the sustainability software that does the actual computation — it does not replace either.
AI agents & workflows
SDAI open infrastructure
Sustainability software
Resources
Docs, examples, and updates to help you integrate and extend the open-source layer.
Issues, pull requests, and method critiques are all welcome. Everything we publish is intended to be inspected, challenged, and reused.
We're interested in how the connectors get used in practice — and what's missing. Tell us what you're integrating.