Programmatic licensing for
AI training, inference, RAG, and fine-tuning.
Per-workload license scope. Cryptographic attestation per event. EU AI Act Article 53 compliant by default. REST API and native MCP server.
Live on Tempo · cryptographic attestation per license event
Scraping just became a legal liability.
Premium content licensing isn't optional for AI labs anymore - it's procurement infrastructure. Opedd is the rail.
NYT v. OpenAI alleged damages. Reddit, Axel Springer, AP, NewsCorp all converging on the same legal theory.
EU AI Act Article 53 - training-data transparency obligations enforceable since 2025. Fines up to €15M or 3% of global turnover.
Publisher × AI-lab licensing deals signed publicly since 2023 - NewsCorp/OpenAI, FT/OpenAI, Axel Springer, AP, Reddit/Google, and more.
Four workloads. Four scopes. One rail.
License only what your stack consumes. Pay only for the events that fire.
Training corpus access
Bulk publisher content licensed for model pre-training and continued pre-training. Per-publisher scope, per-vertical filtering. Attestation JWT per training run, anchored on Tempo.
Inference-time licensing
Per-event licensing for inference calls that consume publisher content - citations, summaries, quoted material. Attestation per inference event, available via REST or MCP server tool calls.
RAG retrieval
Per-query licensing for retrieval-augmented generation. Native MCP server exposes search + fetch tools directly to your inference stack. Every retrieval ships with a cryptographic receipt.
Fine-tuning corpus
Domain-specific publisher content for fine-tuning runs. Per-vertical scoping (geopolitics, defense, finance, scientific, legal). Per-publisher consent verified at license issuance.
Self-serve. Sandbox in 30 seconds.
From your first API call to your first attested license event - no procurement cycle, no sales call.
Browse the catalog
Public API and MCP server expose the full publisher catalog - verticals, content depth, license tiers. No account required.
Get sandbox access
Sign up, get your API key in 30 seconds. Sandbox token unlocks the full catalog and truncated content previews - zero commitment.
Define your license
Pick publishers · pick workloads (TRAIN / INFER / RAG / TUNE) · pick duration. Pay via Stripe. License issued instantly.
Integrate
REST API · Webhooks · MCP server · TypeScript and Python SDKs. Every event ships with a cryptographic receipt and attestation JWT.
High-expertise content where buyer need is sharpest.
The verticals where licensed analyst content materially improves your model's outputs.
Independent analysts, regional-affairs writers, sovereign-risk newsletters.
Sector analysts, credit researchers, niche financial newsletters.
OSINT specialists, military-affairs writers, security analysts.
Domain experts, life-sciences researchers, healthcare-policy writers.
Regulatory analysts, jurisdictional commentary, compliance commentary.
Your compliance posture, on-chain, every event.
EU AI Act Article 53 compliance isn't a quarterly audit deliverable. It's a per-event JWT shipped with every license call. Your legal team gets the receipts they need; your model gets the content it can defend.
- iss:
- opedd.protocol
- art:
- 53
- anchor:
- tempo:0x9f3a…c2e1
- scope:
- inference
- iat:
- 1748566800
- EU AI Act Article 53 attestation JWT issued per license event - regulator-ready training-data provenance.
- Cryptographic receipt for every event - publicly verifiable on Tempo blockchain.
- CDSM Article 4 publisher opt-in verified at license issuance - no scraped material, no opt-out ambiguity.
- Audit trail: every license event timestamped, attested, immutable.
- Per-license terms documented in the attestation payload - scope, duration, sublicensing rights all on-chain.
From zero to attested in 30 seconds.
Sandbox token. Full catalog access. Three integration paths. No procurement cycle.
# pip install opeddfrom opedd import Opedd client = Opedd( buyer_token="opedd_buyer_live_...", # per-article content access_key="ent_...", # enterprise feed) # Stream your licensed feed as NDJSONfor row in client.feed.stream_ndjson(limit=5000): if row["content_access"] == "included": ingest(row["content_body"], metadata=row) else: # metered (filtered-scope) keys ship content_body=None - # fetch full text per article, billed per call article = client.content.get(row["id"]) ingest(article["content_body"], metadata=row)Ship licensed content into your stack.
Sign up. Get your API key. Start the sandbox. No sales cycle, no procurement nightmare.
Need platform-wide coverage or a custom deal? Contact sales
Live on Tempo · cryptographic attestation per license event.