Permanent identity
A governed subject, target, forecaster, and receipt digest identify exactly which prediction is being discussed.
Public infrastructure for verifiable forecasts
Your forecasts. A lasting public record.
Publish forecasts from people, statistical models, and AI in one inspectable library. Give your audience a permanent link to what you predicted, when you predicted it, and the evidence behind it.
No account needed to browseOne receipt per forecastOptional per-receipt blockchain proof
Where Forecast Library began
The original Open Forecast Receipt demo: a forecast, a portable record, and an independent integrity check.
Watch on YouTube376
forecast subjects
4,511
public receipts
8
forecaster profiles
60
verified proofs
Live public catalog counts. Proof status is reported without implying accuracy.
One forecast. One digest. An independently inspectable anchor.
BLOCKCHAIN PROOF
Evidence your audience can check beyond this website.
Choose an optional public blockchain anchor when submitting your forecast receipt. Anyone can compare the downloaded record with its onchain digest and inspect the attestation, issuing address, transaction, and anchoring time.
- Seal the record. Canonical JSON and SHA-256 produce a fingerprint of the forecast payload.
- Anchor the fingerprint. Ethereum Attestation Service (EAS) records a compact forecast projection and digest. For each iPulse AI Showcase asset and batch, all included advisor forecasts are submitted together in one transaction. The number of forecasts can vary by asset and batch. Each receives an individual attestation ID within that submission.
- Verify independently. Download the full receipt, recompute its digest, and compare it with the public chain record.
The integration supports Base mainnet and Base Sepolia testnet; every issued proof identifies its actual network. A receipt is only labelled verified when its onchain proof has been checked. Retrospective anchoring proves existence by the later anchoring time, not by the original forecast date. It does not prove accuracy or authorship.
Request blockchain proof with your submissionOpen-source origins. A shared research record.
Every forecasting method deserves an inspectable record.
Forecast Library grew out of Open Forecast Receipt, an open-source initiative to make a prediction portable, inspectable, and independently verifiable. Its scope is forecasting science across domains, from markets and weather to demand and scientific research, including human, quantitative, and AI forecasts. iPulse AI is a separate financial-market research product and our first live publisher. Other domains will need their own reviewed receipt profiles.
Human judgment, statistical models, ensembles, and AI all contribute to forecasting. We expect increasingly capable AI to play a larger role, but the Library is designed around the forecast and its evidence, independent of who or what produced it. Our ambition is a shared record that people and research agents can both use.
That vision includes Agent-to-Agent (A2A) collaboration and other agentic workflows. Today, developers can use the open receipt schema, deterministic verifier, and public receipt JSON API. Automated agent integrations and broader evaluation workflows are the next direction.
The real-world problem
Make your original prediction easy to find and check.
Reports get revised. Dashboards change. Forecasts disappear into posts and spreadsheets. Preserve the original prediction, method, assumptions, and timing so your audience can return to the same record later.
A permanent receipt gives your forecast a reference your audience can cite, download, and independently verify.
Documentation does not make a forecast correct. It makes honest reconstruction, comparison, and evaluation possible.
A complete time map
Creation, knowledge boundaries, supplied context, publication, maturity, and evaluation remain separate dates.
Versioned provenance
Preserve the method, assumptions, source evidence, and relevant model or code versions behind a forecast.
Integrity and corrections
Canonical JSON produces a deterministic digest. Corrections are append-only records instead of silent rewrites.
From prediction to evidence
One forecast. One receipt. A record that survives.
1
Create
Map the subject, target, forecaster, predicted values, time map, and provenance into the open format.
2
Seal and publish
Canonicalize the receipt, compute its SHA-256 digest, and publish the browse projections.
3
Prove (optional)
Anchor that individual receipt digest onchain and link the returned attestation without rewriting it.
4
Evaluate later
After the horizon matures, append the governed outcome and evaluation while preserving the original.
Evidence, ownership, and limits
Built in public. Operated by a real organization.
Future Edge Group FZE operates the Library and publishes the open-source receipt standard. iPulse AI is the first production client, contributing a real public collection rather than a hand-picked mock demonstration.
Accountable operator
Future Edge Group FZE · Formation 4414073
Open implementation
Public schema, examples, verifier, and tests
Explicit limitations
Integrity and timing are not accuracy or truth
Designed for inspection
Useful to publishers, reviewers, developers, and the public.
Publish your track record
Give clients and readers a permanent link to your original prediction, whether you forecast personally, with statistical models, or with AI.
Make research reproducible
Keep the original values, dates, assumptions, and evidence available for later comparison with outcomes.
Build on an open format
Download structured JSON and verify its digest with an open-source toolkit. Keep your integration independent of page layouts.
Show your evidence
Let anyone inspect your public record without an account. Request independent blockchain anchoring when extra traceability matters.
Open format. Independent verification.
A structured forecast record you can keep and reuse.
What to send
Contact support with your forecast values, subject, target, dates, method, and public evidence. Use our submission checklist or provide an OFR JSON receipt. Support reviews publication rights and format compatibility before listing.
The current receipt format
Open Forecast Receipt v0.1.0 uses JSON Schema, RFC 8785 canonicalization, and a SHA-256 payload digest. The live profile covers financial forecasts; human and statistical submissions are welcome for review, while other domains may require a new versioned profile.
What is preserved
The receipt contains one forecast and its declared provenance. It is not an AI chat-history, agent-memory, or hidden chain-of-thought upload. Permanent forecast IDs and digest-addressed JSON remain stable when navigation or presentation changes.
Frequently asked questions
Forecast Library, clearly explained.
Direct answers about receipts, verification, blockchain proof, evaluation, and governance.
What is Forecast Library?+
Forecast Library is a public, browsable library of forecast subjects, forecasters, forecasts, receipts, integrity checks, optional blockchain proofs, and later evaluations. It is built on the Open Forecast Receipt standard.
What is an Open Forecast Receipt?+
An Open Forecast Receipt is a portable, structured record of one forecast. It preserves what was predicted, who or what produced it, when it was generated, the relevant knowledge and input-context boundaries, provenance, and a deterministic integrity digest.
What does blockchain proof establish?+
An optional blockchain attestation can establish that a particular receipt digest existed by an anchoring time and has not changed. It does not prove that the forecast is accurate, truthful, unbiased, or based on good reasoning.
Does every forecast need to be stored on a blockchain?+
No. Every receipt can be verified locally from its canonical data. Blockchain anchoring is an optional, independent proof layer for selected receipts.
Is this only for AI or financial forecasts?+
No. Markets and AI forecasters are the first live use case. The subject–target–forecaster–forecast–receipt structure can also describe human, quantitative, ensemble, and hybrid forecasts in other domains.
When is a forecast evaluated?+
Only after its defined horizon or target has matured. Evaluation is appended as a separate record so the original forecast remains unchanged.
Can a receipt be corrected?+
Yes, but never silently. A correction receives its own identity and digest and references the record it corrects, preserving the audit trail.
Who operates the Library and how can I submit?+
Forecast Library is operated by Future Edge Group FZE and first used by iPulse AI. During this phase, public submissions are coordinated by email at support@ipulseai.com; no account or payment is required.
How does Forecast Library relate to agentic workflows?+
Forecast Library grew out of Open Forecast Receipt, an open-source initiative for portable, verifiable forecast records. Our ambition is for research agents to exchange receipts, preserve provenance across handoffs, and append evaluations when outcomes become known, including through Agent-to-Agent (A2A) workflows. The open schema, verifier, public library, and receipt JSON API are available today; automated agent integrations are a future direction.
Start with the public record
Inspect the forecasts. Recompute the receipts. Judge the evidence.
Public browsing needs no account. To propose a public submission during this phase, contact the Library team directly.
