What shipped.

Newest at top.

v1.3 · Local-attest + content 2026-04-26

Local model attestations + 50 SEO guides.

  • New /v1/store-run endpoint accepts pre-signed run JSON from any local runner
  • Open-source _local_runner.py against Ollama (Windows + WSL), same canonical sample sets as cloud runs, fully comparable
  • 78 attested runs across 6 local open-weight models, qwen3.6:35B, qwen3.6:27b-dense, mistral:7B, deepseek-coder-v2:15.7B, llama3:8B, glm-4.7-flash:30B
  • New /local dashboard, model leaderboard + cross-source comparison per benchmark
  • Trust ladder explicit: Verified⛓ Attested Local Self-reported, leaderboards now sort trust-tier-first
  • 50 long-form benchmark guides at /articles with FAQPage + Article structured data
  • 109 self-reported listings seeded from public vendor announcements
  • Sitemap extended to 377 URLs
v1.2 · Inline worker 2026-04-25

Pure-JS inline cron worker.

  • /api/cron/drain processes managed-mode runs every 2 min, Anthropic API key fallback, no GitHub Actions dependency
  • Stuck-job reaper: any run older than 120s in :processing gets refunded + marked failed
  • 24 simple-grade benchmarks runnable inline (numeric + letter + text scoring)
  • Aligned Layer ZK proof requests honestly recorded as verification.status: 'pending', queued at bl:queue:aligned for operator
v1.0 · Launch 2026-04-22

Public launch.

  • 77 services across 16 categories, 30 benchmark suites
  • 62 attested runs, 30 via Aligned Layer ZK proofs
  • Proof viewer with Ethereum block + tx explorer links
  • Compare up to 4 services side-by-side
  • Open API: all data readable at /api/*.json
  • Open-source runner (benchlist-runner)
  • Dispute system with 0.1 ETH bond + 3-attestor quorum
  • MIT-licensed SDKs: Python, Node, Go
v0.9 · Private beta 2026-04-19

Attestor registry + 30 seed services.

  • Attestor operator spec, Ed25519 key rotation
  • Schema v1 frozen for service/benchmark/run
  • Reference runner rewritten in Rust for SP1 circuit size
  • 30 services seeded across LLM, memory, code-agent categories
v0.5 · Testnet 2026-04-10

First on-chain attestation.

  • Aligned Layer Holesky integration end-to-end
  • First run attested: LongMemEval · 94.6%
  • Verifier contract deployed at 0xeF2A…606c
v0.1 · Prototype 2026-04-01

First commit.

  • Project scoped, data model v0, marketplace page live