Evidence over narrative
Every score traces back to source text. Where evidence is weak, the tool says so and labels the level rather than smoothing it into a number.
About
Rigorous impact
Rigorous impact
measurement should
not be a luxury.
Serious impact measurement has historically meant a consultant, a six-figure budget and a three-month engagement. That is affordable for a large institutional fund and out of reach for almost everyone else — which is precisely backwards, because the funds with the least measurement capacity are often closest to the outcomes.
Impact Vision exists to close that gap. It embeds the frameworks the sector already agreed on — IRIS+, the SDG targets, the five dimensions of impact, NESTA's standards of evidence, and the disclosure regimes now arriving in law — into a tool that a two-person fund can run on a laptop, for free, and read the source of.
What we hold to
Six commitments that shape what goes into the tool and what stays out of it.
Open by default
MIT licensed, all of it. The mappings, thresholds and weightings are readable in the repository, because a scoring engine you cannot inspect is just an opinion with a logo.
Standards, not invention
IRIS+, the SDG targets, the five dimensions, NESTA, SFDR, CSRD, ISSB, PCAF, TNFD. The tool applies existing discipline faster; it does not invent a proprietary rating.
Useful to both sides
Investors get diligence they can defend. Companies get a clear list of what to measure and why — through the investee portal — rather than an annual surprise.
Local first
SQLite, JSON catalogues, self-contained HTML reports. Point it at a local Ollama model and no company data leaves the machine.
Honest about limits
Proposed, pending and beta standards are labelled as such. The tool assesses readiness and alignment; it is not assurance, and it is not legal advice.
The wider problem
Capital is moving.
Capital is moving.
Verification is not
keeping up.
Trillions now carry a sustainability label. The infrastructure for checking those labels — consistent metrics, honest counterfactuals, evidence grading, comparable portfolio roll-ups — remains thin, manual and expensive.
Regulators have noticed. SFDR, CSRD, the ISSB standards, the UK FCA anti-greenwashing rule and the EU Empowering Consumers Directive all point the same way: a claim now has to be substantiated, and the substantiation has to be documented.
The uncomfortable implication for the sector is that a lot of existing impact reporting will not survive that test. The optimistic one is that the tooling to fix it is not hard to build — it just has to be built openly, so that a fund can inspect the method rather than trust a vendor.
For fund managers
Impact diligence that stands up in an IC and in an LP audit, produced in the time a first screen actually gets.
For portfolio companies
A clear, standards-based list of what to measure — set once, at investment, instead of guessed at annually.
For researchers & students
A working, readable implementation of the major frameworks. Fork it, break it, cite it.
Roadmap
Where it is going.
Development runs in waves, each documented in the repository with its own plan and changelog entry.
Wave v2
Institutional readinessData contracts, investee data collection, climate accounting, LP reporting, assurance, causal impact and governed AI.
Wave v3
Trust infrastructureEvidence governance, stakeholder voice, verification workspace, LP narrative from verified data only, portfolio natural-language query. Shipped.
Wave v4
Consultant engagement suiteProposal, data-room and training workflows for advisers. Backend shipped; frontend and paid-data wiring deferred.
Wave v5
Legally current & frontier measurementRegulatory deltas through 2026, monetary valuation, welfare quantification, LCA/LCSA. Shipped.
Wave v6
Comparable, assured, connectedComparability across portfolios, assurance bundles and external connectivity. Planning baseline with initial surfaces shipped.
Contribute
Ways in.
The most valuable contributions are rarely code. A framework mapping corrected by someone who works with that standard daily is worth more than a refactor.
- Correct or extend a framework mapping — 59 cross-referenced concepts, all of them improvable.
- Add sector packs to the DD checklist for sectors you know well.
- Report where a score felt wrong, with the input that produced it.
- Contribute benchmark data, or point to public datasets worth ingesting.
- Improve the reports — they are self-contained HTML and easy to edit.
Licence
MIT
Use it in commercial fund work, fork it, white-label the reports, ship it internally. Attribution is the only ask.
Built on
OpenHarness
The agent framework underneath the tool router, skills and permission model. HKUDS/OpenHarness
Acknowledgments
Standing on other people's work
AvantFaire Investment Management, GIIN, Pacific Community Ventures, Seraf, the Impact Management Project and OpenHarness — for the frameworks, practices and infrastructure that inform Impact Vision.
Important. Impact Vision is a measurement and assessment tool. It is not investment advice, not a regulatory filing, not third-party assurance, and not a substitute for professional judgment or legal counsel. Framework names and content belong to their respective standard-setters. Outputs should be reviewed by a qualified person before they inform any investment or disclosure decision.
Free · MIT · no account
Start with one deck.
Clone it, run it against a company you already know well, and see whether the assessment matches your instinct. That is the honest test.