Summary
This summary is here to be read. It is not part of the agreement, and if it and the articles below ever disagree, the articles win.
- open-climate.ai gives you a searchable, harmonised database of emission factors and a set of tools that map your own data (procurement lines, activity data, files) onto those factors.
- The data you send us stays yours. We do not sell it, we do not show it to other customers, and we do not use it to train large language models.
- We do use it, in a restricted way set out in article 7, to make our matching and search algorithms better.
- Your data is hosted in the European Union and stays there.
- The emission factors come from their publishers under the publishers' licences. Some are open, some are share-alike, some may not be redistributed at all. Article 8 tells you which obligations travel with the numbers you export.
- The results are estimates produced in part by AI systems. They are a starting point for a greenhouse gas inventory, not an audited one. You remain responsible for what you report.
1. Who we are, and what these terms cover
1.1 The service described below is operated by Songbird AI BV, a private limited company (besloten vennootschap) incorporated under Belgian law, with registered office at Rue Emile Wittmann 50, 1030 Schaerbeek, Belgium, registered in the Crossroads Bank for Enterprises under enterprise number 1033.450.866, VAT BE 1033.450.866 ("we", "us", "Songbird AI").
1.1.1 open-climate.ai is a trading name and a brand of Songbird AI BV. Where these Terms, the website or an invoice say open-climate.ai, the contracting party is Songbird AI BV.
1.2 These Terms of Service (the "Terms") govern all access to and use of the open-climate.ai website, factor explorer, documentation, APIs, MCP server and related tools (together, the "Service"), whether on a free or a paid plan.
1.3 The Service is offered to businesses, public bodies and other professional users only. It is not offered to consumers, and by using it you confirm that you are acting for purposes relating to your trade, business, craft or profession.
1.4 By creating an account, issuing an API key, connecting the MCP server or otherwise using the Service, you accept these Terms on behalf of the organisation you represent (the "Customer", "you"), and you confirm that you are authorised to do so.
1.5 If we have signed a separate written agreement or order form with you, that document prevails over these Terms to the extent the two conflict. Where a third-party data licence conflicts with these Terms, the data licence prevails for the data it covers (see article 8).
1.6 We may publish additional documents that form part of these Terms by reference: the Privacy Policy, the Data Processing Agreement, the Acceptable Use rules referenced in article 5, and the published plan and pricing page. They are linked from the footer of every page on https://open-climate.ai.
2. Definitions
- Customer Data: anything you or your Users send to the Service, including uploaded files, procurement or activity lines, descriptions, quantities, supplier names, prompts and API request bodies.
- Results: the output the Service returns to you, including matched emission factors, calculated emissions, confidence signals, explanations, and exported files.
- Factor Data: the emission factors, their metadata, identifiers, units, system boundaries, licence fields and attribution lines served by the Service, whether sourced from a third-party publisher or produced by us.
- Publisher: the body that issues an emission factor library, for example ADEME, DESNZ, the U.S. EPA or the BBSR.
- User: a natural person authorised by you to use the Service under your account, including through an API key or an MCP client.
- Subscription Term: the period stated in your plan or order form.
3. The Service
3.1 The Service currently comprises, depending on your plan:
- a factor explorer, for searching and reading emission factors;
- exports of selected factors, subject to the licence rules in article 8;
- an HTTP API, and an MCP server for use from AI assistants and agents;
- mapping tools that take your files or line items and return matched factors and calculated emissions;
- documentation and reference material.
3.2 Reading the public factor explorer does not require an account. Exports, the API, the MCP server and the mapping tools do.
3.3 For as long as your subscription runs, you may use the Service and the Results for your own business: to build and maintain your greenhouse gas inventory, and to report it to your regulators, auditors, customers and investors. That right is yours alone, it cannot be transferred or sublicensed, and it ends when the subscription does. Anything not granted here is reserved, and the licence conditions attached to Factor Data apply on top of it (article 8).
3.4 The Service evolves. We may add, change or withdraw features. Where a change materially reduces the functionality of a paid plan, we will tell you at least 30 days in advance, and you may terminate the affected plan with effect from the change date.
3.5 Features marked beta, preview or experimental are provided as they are, may change without notice, may be withdrawn, and are excluded from any service level or availability commitment.
4. Accounts, keys and security
4.1 You must give accurate account information and keep it current.
4.2 API keys, access tokens and MCP credentials identify you. Keep them secret, do not embed them in client-side code or public repositories, and rotate them if they may have been exposed. Everything done with your credentials is treated as done by you.
4.3 Tell us without undue delay at hello@open-climate.ai if you suspect a credential has leaked or an account has been compromised.
4.4 You are responsible for your Users' compliance with these Terms.
4.5 We may suspend an account or a key immediately where we reasonably believe it is compromised, is being used in breach of article 5, or is placing the Service at risk. We will tell you why, and restore access once the cause is resolved.
5. Acceptable use
5.1 You may not use the Service to break the law, infringe anyone's rights, or transmit malicious code.
5.2 The export features are the supported way to take Factor Data out of the Service, and the licence fields travel with the rows when you use them. Taking the data out any other way is not allowed: that covers working around a rate limit or a plan quota, driving the explorer with a browser automation, walking the API to reconstruct the catalogue, and any other systematic extraction, whatever the technique and however slowly it is done.
5.3 You may not resell, sublicense, rent out or otherwise make the Service or the Factor Data available to third parties as a service of your own, unless we have agreed that in writing. Passing individual factors or Results to your own clients as part of your own advisory or reporting work is allowed, within the publisher licence conditions in article 8.
5.4 You may not use the Service, the Factor Data or the Results to train, fine tune, distil or benchmark a machine learning model whose purpose is to match, classify or estimate emission factors, or to build a product that substitutes for the Service. This restriction does not limit what the publisher's own open licence already permits you to do with data you have obtained under that licence.
5.5 You may not misrepresent the Results, present them as verified or audited figures when they are not, or suggest that we or a Publisher endorse your figures, your product or your organisation.
5.6 Fair use. Free and metered plans carry published quotas. We may throttle or suspend usage that materially exceeds them or that degrades the Service for others, after notice where the circumstances allow.
5.7 Agents and the MCP server. When you connect an AI assistant or an autonomous agent to the Service, whether over MCP or the API, it acts under your credentials and on your behalf, and article 4.2 applies to what it does. Give an agent only the access it needs, keep a human in the loop for anything you will report, and remember that the quota it consumes is yours. The MCP connection is for your own use and your own agents. Putting it in front of third parties, or building it into a product you distribute, needs our written agreement under article 5.3.
5.8 No reverse engineering. You may not decompile, disassemble or reverse engineer any part of the Service, and you may not attempt to derive its source code, its prompts, its model weights, its thresholds, its ranking formulas or the internal structure of the Factor Data, whether by inspecting what we send you, by analysing Results, or by any other route. Two things sit outside this restriction. The first is the rights you hold under Articles 5(3) and 6 of Directive 2009/24/EC, in Belgium Articles XI.299 and XI.300 of the Code of Economic Law, which we neither restrict nor purport to restrict, save that where you need interoperability information you must ask us for it in writing and allow us a reasonable period to supply it before acting for yourself. The second is our open source components, which are governed by their own licences (article 8.1.c) and by nothing in this article.
5.9 No model extraction. You may not query the Service in a manner designed to reconstruct how it works rather than to obtain an answer you need. That covers systematic or adversarial querying intended to infer, replicate or approximate a model, a classifier, a ranking, a threshold or a matching rule, and it covers assembling Results into a training set, a label set or a lookup table that stands in for the Service. Ordinary use, including heavy use within your plan, is not caught by this article. What is caught is querying whose purpose is reconstruction rather than use.
5.10 No competing use. You may not access the Service for the benefit of a person who offers, or is developing, a product that competes with it, and you may not use the Service or the Results to assemble a competing database of activity to factor mappings. You may not publish a benchmark or a comparison of the Service without our written consent, which we will not unreasonably withhold where you disclose the methodology to us first and give us an opportunity to comment. None of this restricts your own internal evaluation of the Service, and none of it restricts what a Publisher's own licence already permits you to do with data you obtain from that Publisher.
5.11 No unauthorised security testing. Penetration testing, vulnerability scanning and load testing against the Service require our written authorisation in advance. If you discover a vulnerability, report it to hello@open-climate.ai and give us a reasonable period to remedy it before disclosing it to anyone else. We will not pursue a good faith reporter who stays within those limits.
5.12 Traceability of exports. Exports and API responses may carry markers that associate a copy with the account it was issued to. You may not remove, alter or obscure them, and we may use them to establish the provenance of an unauthorised copy.
6. Plans, fees and payment
6.1 Plans, quotas and prices are those published at https://open-climate.ai/pricing/ or stated in your order form.
6.2 Unless stated otherwise, fees are exclusive of VAT and other taxes, are invoiced in advance for the Subscription Term, and are payable within 30 days of invoice date.
6.3 Subscriptions renew automatically for successive terms of the same length unless either party gives notice under article 12.
6.4 We may change prices for a renewal term with at least 60 days' notice before the renewal date. If you do not accept the new price, you may terminate with effect from the end of the current term.
6.5 Late payment: we may charge statutory interest and, after written warning and a further 14 days, suspend the Service.
6.6 Fees paid are not refundable except where these Terms or mandatory law say otherwise.
7. Your data
7.1 It stays yours
You own Customer Data. Nothing in these Terms transfers ownership of it to us. You grant us only the rights set out in this article, and only for as long as we need them.
7.2 What we use it for
We process Customer Data to:
a. provide the Service to you, which includes parsing your files, classifying and matching your lines, calculating emissions and returning Results; b. operate, secure, monitor and troubleshoot the Service, including keeping the logs and job records needed to explain a given Result to you; c. meet our legal obligations; d. improve the Service, within the limits of articles 7.3 and 7.4.
We do not sell Customer Data. We do not share it with other customers. We do not disclose it to a Publisher or to any third party except the sub-processors listed under article 7.7, and except where the law compels us.
7.3 We do not train language models on your data
We do not use Customer Data or Results to train, fine tune or otherwise adapt large language models, whether our own or a third party's.
Where the Service calls a third-party model provider to process your data, we do so under terms that contractually prohibit that provider from using the content of our requests to train or improve its models. The current list of model providers is published with our sub-processor list.
7.4 We do use it to improve the algorithms
You agree that we may use Customer Data and Results to evaluate and improve the non-generative parts of the Service: retrieval, search ranking, classification, unit conversion, the harmonised factor model, quality checks and the accuracy measurement we run against them. In practice that means counting what happened, measuring where the Service got an answer wrong, and building test cases that keep it from getting it wrong again.
This use is subject to all of the following limits:
a. we use the minimum data needed, and we remove or pseudonymise direct identifiers where the improvement purpose does not require them; b. the outcome is statistics, evaluation metrics, rules, vocabulary entries and test cases. We may keep extracts of Customer Data as evaluation fixtures for as long as the fixture is useful, so that a fault we fixed for you stays fixed. A fixture is de-identified before it is kept, and it is used to test the Service, never to answer another customer's query; c. no output of this work identifies you, your suppliers or your volumes to any other customer or to the public; d. it never includes training a large language model (article 7.3); e. we do not publish, sell or license the underlying Customer Data.
If you would rather we did not use your data this way at all, tell us at hello@open-climate.ai and we will exclude your account.
7.5 Your data stays in Europe
Customer Data and Results are stored and processed on infrastructure located in the European Union. Our application servers, our databases, our object storage and our backups are in EU data centres.
Where the Service uses a third-party model provider or other sub-processor to process Customer Data, we use endpoints and configurations that keep that processing inside the European Union or the European Economic Area. If a sub-processor cannot be operated that way, we do not send it Customer Data.
Our own staff and contractors access Customer Data from within the EU or the EEA, and only where needed for support, security or the purposes in article 7.2.
7.6 Retention and deletion
We keep Customer Data, Results and the associated job records for as long as your account is open, so that you can re-open, re-export or audit a past calculation, and so that we can investigate a disputed figure. Nothing expires on a timer while you are using the Service.
You may ask us at any time to delete a file, a job or your whole account's data, and we will do so within 30 days, except for:
a. data we must keep to comply with a legal obligation, for example invoicing records; b. backups, which expire on their own cycle of at most 90 days; c. the aggregate and pseudonymised outputs described in article 7.4, which cannot be traced back to you and are not deleted.
When your account closes, article 12.4 gives you 30 days to export. After that we delete Customer Data and Results within 90 days, subject to the same exceptions.
7.7 Personal data
To the extent Customer Data contains personal data, you are the controller and we are the processor. Our Data Processing Agreement at https://open-climate.ai/legal/dpa/ governs that processing, and it includes the current list of sub-processors, the security measures we apply, and how we handle data subject requests and breach notification. It forms part of these Terms.
We process account and billing data about your Users as a controller in our own right. The Privacy Policy at https://open-climate.ai/privacy/ explains that.
7.8 Confidentiality
Each party will keep the other's non-public information confidential, use it only to perform these Terms, and protect it with at least the care it applies to its own confidential information. This survives termination by 3 years, and indefinitely for anything that is a trade secret. Customer Data is your confidential information.
8. Emission factor data, and the licences that come with it
This is the article to read before you export anything.
8.1 Three layers
The Factor Data you see has three layers, and they are licensed differently:
a. The publishers' data. The numbers, and the descriptions that came with them, belong to the Publisher that issued them. Importing a library, cleaning it and serving it gives us no ownership of it, and we cannot pass on rights we were never granted: whatever the Publisher allows is the ceiling on what you may do. b. Our layer. The harmonised data model, the stable identifiers, the translations, the classifications, the unit and boundary normalisation, the embeddings, the crosswalks between libraries and the search behaviour are ours. We license them to you under article 3.3 for the Subscription Term. c. Our open source components. Parts of the platform are published under open source licences. Those licences govern that code, and nothing in these Terms restricts a right they grant you.
8.2 Licence tiers
Every library release carries a licence tier, and the tier decides what the Service serves you:
| Tier | What you get | What it means |
|---|---|---|
| open | The factor and its values in full | Public domain or attribution only. Redistribution is granted. |
| copyleft | The factor and its values in full | Redistribution is granted, but your copies must carry the same licence. |
| restricted | The description, with values withheld | No redistribution right has been established, so you can see that a factor exists but not its value. |
| prohibited | No values at all | The publisher's terms forbid redistribution. |
The tier, the licence code, the licence URL and the attribution line are fields on every factor, and they travel on every row you export.
8.3 Your obligations when you use or pass on factors
a. Attribution. Reproduce the attribution_text of every factor you use, as
written, in any report, file or product that carries the value.
b. Share alike. Where a factor is copyleft, anything you distribute that
contains its values must carry the same licence. A file that mixes libraries
has to satisfy every licence in it.
c. No implied endorsement. Crediting a Publisher is required. Suggesting
that a Publisher endorses you, your product or your figures is not allowed,
and some licences say so explicitly.
d. Restricted and prohibited factors. Cite them by identifier and library,
and obtain the values from the Publisher under your own licence with them.
Do not attempt to recover a withheld value from the Service.
e. Modifications. Where a licence requires you to state what you changed,
state it. Where a licence requires values to be passed on unchanged, do not
adjust them and then attribute the result to the Publisher.
8.4 What we do not warrant about licences
We record licence terms in good faith and keep them current, and the licence fields are part of what we consider the product. They are not legal advice. A Publisher's own terms, linked from every library page, are what bind you, and where our record and the Publisher's terms disagree, the Publisher's terms win. Some answers are recorded as unclear, and where they are, you should ask the Publisher before relying on that use.
8.5 Changes over time
Publishers revise their libraries and occasionally their terms. A tier is decided per release: a Publisher that changes its terms changes the tier of its next release, not of the releases you already used. Where a Publisher withdraws our right to distribute a library, we may remove it from the Service, and we will tell you.
Factor values change between releases. Where the Service offers a way to pin a release, use it if your reporting needs stable numbers across a period.
9. Results, accuracy, and AI
9.1 The Results are estimates. They are produced by matching your descriptions to published factors and multiplying by quantities. They are not measurements, they are not an audit, and they are not a certification.
9.2 We build the Service to be useful for greenhouse gas reporting and to be traceable back to a source, but we do not warrant that any Result is accurate, complete, or fit for a particular regulatory framework, reporting standard or assurance engagement. Deciding whether a factor is the right one for your activity is your decision, and yours to defend.
9.3 Parts of the Service use AI systems, including large language models, to classify, match and explain. Their output is probabilistic. It can be wrong, and it can be wrong while looking confident. You must apply human review proportionate to the materiality of what you are reporting, and you must not present an unreviewed Result as a verified figure.
9.4 The AI features of the Service search, classify and explain reference data. They are not designed for, and are not offered for, any of the uses listed in Annex III of Regulation (EU) 2024/1689 (the AI Act), and they take no decision about a person. Article 9.3 is how we meet our transparency duty: you are being told, plainly, which parts of the output a model produced and that it can be wrong. If you intend to embed the Service in a system that would be high risk under that Regulation, tell us before you do, because the obligations that follow fall on you as the deployer and some of them need our cooperation.
9.5 Where the Service reports a confidence signal, an assumption, or an unmatched line, that signal is part of the Result. Suppressing it when you pass Results on to others is your responsibility, not ours.
10. Availability and support
10.1 We aim for high availability but, unless an order form states a service level, the Service is provided without an availability commitment.
10.2 We may take the Service down for maintenance. We will schedule planned maintenance outside European business hours where we reasonably can, and announce it in advance.
10.3 Support channels and response targets, if any, are those stated in your plan.
11. Intellectual property
11.1 We and our licensors retain all intellectual property rights in the Service, in our layer of the Factor Data (article 8.1.b), in the documentation and in our brand. Nothing in these Terms transfers any of it to you beyond the licence in article 3.3.
11.2 You retain all rights in Customer Data.
11.3 You own the Results as they apply to your business, subject to the Publisher licence conditions in article 8 and to the restrictions in article 5.
11.4 If you send us feedback, suggestions or bug reports, we may use them without restriction and without owing you anything. Feedback is not Customer Data.
11.5 You may state that you use open-climate.ai. Any other use of our name or logo needs our written consent, and we will not use yours as a reference without your consent.
11.6 Database right. Obtaining the libraries, verifying them, harmonising them onto one model, translating and classifying them, normalising their units and boundaries and presenting them as one searchable whole represents a substantial investment by us. We assert the sui generis database right under Directive 96/9/EC, in Belgium Book XI, Title 7 of the Code of Economic Law, in that database. Extraction or re-utilisation of a substantial part of it, and repeated or systematic extraction or re-utilisation of insubstantial parts, are prohibited except as these Terms allow. This right protects our compilation and the investment behind it. It gives us no rights in the underlying facts, and it takes nothing away from what a Publisher's own licence grants you in that Publisher's data obtained from that Publisher.
11.7 Trade secrets. The matching pipeline, the prompts and prompt chains, the ranking and scoring logic, the classification taxonomies, the crosswalks between libraries, the evaluation sets and the accuracy measurements behind the Service are trade secrets within the meaning of Directive (EU) 2016/943, in Belgium Book XI, Title 8/1 of the Code of Economic Law. They are our confidential information under article 7.8, we take reasonable steps to keep them secret, and the remedies for unlawful acquisition, use or disclosure under that Directive apply in addition to anything in these Terms.
12. Term, suspension and termination
12.1 These Terms apply from your first use of the Service until all your subscriptions have ended and your account is closed.
12.2 Either party may terminate a subscription with 30 days' notice before the end of the then current term. Free plans may be terminated by either party at any time.
12.3 Either party may terminate immediately on written notice if the other commits a material breach and does not cure it within 30 days of being asked to, or becomes insolvent. We may also terminate immediately for a breach of article 5 that we reasonably consider serious, or for payment default that persists after article 6.5.
12.4 On termination: your access ends, you must stop using the Service, and you have 30 days to export your data. After that we delete Customer Data and Results in line with article 7.6. You may keep Results you have already exported and use them for your own archival, audit, statutory and internal monitoring purposes, subject to the Publisher licence conditions, but no further rights are granted.
12.5 Articles 5.4, 5.8 to 5.12, 7.1, 7.8, 8.3, 11, 13, 14, 15 and 18, and any accrued payment obligation, survive termination.
13. Warranties, and the limits of them
13.1 Each party warrants that it has the authority to enter into these Terms.
13.2 We warrant that we will provide the Service with reasonable skill and care.
13.3 Beyond article 13.2, and to the fullest extent the law allows, the Service, the Factor Data and the Results are provided as they are, without any warranty of merchantability, fitness for a particular purpose, accuracy, completeness or non-infringement.
13.4 You warrant that you have the right to send us the Customer Data you send, including any personal data and any third-party confidential information in it.
14. Liability
14.1 Neither party limits its liability for death or personal injury caused by its negligence, for fraud or fraudulent misrepresentation, or for anything else that cannot be limited under applicable law.
14.2 Neither party is liable for indirect or consequential loss, loss of profit, loss of revenue, loss of anticipated savings, loss of goodwill, or the cost of regulatory penalties incurred by the other, even if the loss was foreseeable.
14.3 Subject to 14.1, our total aggregate liability arising out of or in connection with these Terms is capped at the fees you paid us in the 12 months before the event giving rise to the claim, or EUR 1,000 where you are on a free plan.
14.4 We are not liable for a decision you take, a figure you report, or a penalty you incur on the basis of a Result, or for the accuracy of data supplied by a Publisher.
14.5 You must bring any claim within 12 months of becoming aware of the facts giving rise to it.
15. Indemnity
15.1 You will indemnify us against third-party claims arising from Customer Data you had no right to send us, or from your use of the Service in breach of article 5 or article 8.3.
15.2 We will defend you against a third-party claim that the Service, as provided by us and used in accordance with these Terms, infringes that third party's intellectual property rights, and we will pay the damages finally awarded. This does not cover claims arising from Customer Data, from a Publisher's own data, or from use in breach of these Terms.
16. Changes to these Terms
16.1 We may change these Terms. For a material change affecting a paid plan we will give at least 30 days' notice by email to your account contact and by a notice on the site.
16.2 If you do not accept a material change, you may terminate the affected subscription with effect from the date the change takes effect, and we will refund any prepaid fees for the unused period.
16.3 Continuing to use the Service after a change takes effect means you accept it. We keep previous versions of these Terms and will send you any of them on request to hello@open-climate.ai.
17. General
17.1 Sub-contracting. We may use sub-processors and sub-contractors, and we remain responsible for their performance. Article 7.7 governs sub-processors that handle personal data.
17.2 Assignment. Neither party may assign these Terms without the other's consent, except to a successor of its business, on notice.
17.3 Force majeure. Neither party is liable for a failure caused by an event beyond its reasonable control, for as long as the event lasts.
17.4 Severability. If a provision is unenforceable, the rest stands and the provision is read down to the minimum extent needed.
17.5 No waiver. Not enforcing a right is not a waiver of it.
17.6 Entire agreement. These Terms, the documents they incorporate, and any order form are the whole agreement between us on their subject matter, and replace any earlier understanding.
17.7 Notices. Notices to us go to hello@open-climate.ai and to the registered address above. Notices to you go to your account contact email.
17.8 Language. These Terms are drafted in English. Where we publish a translation and the versions differ, the English version prevails.
18. Governing law and jurisdiction
18.1 These Terms are governed by Belgian law, excluding its conflict of law rules and the UN Convention on Contracts for the International Sale of Goods.
18.2 The courts of Brussels, Belgium have exclusive jurisdiction, except that either party may seek injunctive relief in any competent court. Which language section of those courts hears a case, and in which language it proceeds, is determined by Belgian law on the use of languages in judicial matters; we do not designate one here.
19. Contact
Songbird AI BV Rue Emile Wittmann 50, 1030 Schaerbeek, Belgium Enterprise and VAT number BE 1033.450.866 Everything, including privacy questions and security reports: hello@open-climate.ai