// CHANGELOG · PUBLIC LEDGER
we ship daily.
Last 50 commits to agentspeech.ai · pulled live from git
[2026-05-26] 34c0e74 Merge polish/live-everything-6-4: POLISH_6_4_LIVE + CLEANUP_6_4_GREEN
[2026-05-26] d2c2ae5 cleanup: 3 stale vitest assertions on copy/asset drift
[2026-05-26] 3858fd5 polish: live-everything pass — 7 component edits + 9 new live pages
[2026-05-26] 8284bee DAY_1B_MANIFESTO_PORTED: chip-locked layout, handshake background, Claude Design copy verbatim
[2026-05-26] d54300c day 1 phase 1E: agentspeech.ai apex + www LIVE with LE TLS
[2026-05-26] 9563744 day 1 phase 1D: seed receipts wall with 10 verified entries (5 Monty UUIDs HTTP-200 audited + 5 Mike infra milestones, sourced from publish_audit.jsonl, no fabrication)
[2026-05-26] 8a21dc7 day 1 phase 1C: wire DeepSeek + x402 to /api/v1/price-suggest (5/5 tests pass)
[2026-05-26] 754b8d6 day 1 phase 1B: apply truth audit — 24 rephrases + 3 deletes across 13 files
[2026-05-26] 569c4f0 day 1 phase 1A: full-site truth audit (41 claims surveyed, 30 false/imprecise)
[2026-05-26] 4f07298 pre-day1: SIWE + /claim hardening before truth audit
[2026-05-25] 05d1c22 phase 8R-7 D: PageSpeed Insights validation passes — all 3 pages >=85 all axes
[2026-05-25] 53c8ed7 phase 8R-7 B1-B5: Caddy public exposure at staging.agentspeech.ai (LE TLS, prod systemd rebuilt, Lane A live)
[2026-05-25] 2076277 phase 8R-7 C2 + D1: cred regen manifest + verify helper + laptop Lighthouse runner
[2026-05-25] c7e8e68 phase 8R-7 A4: Footer split RSC + FeedbackWidget island (TBT target)
[2026-05-25] f622ac4 phase 8R-7 A3: 1149-*.js audited — Next15+React19 framework, not trimmable from app code (Sentry already lazy in 6264 chunk from 8R-1)
[2026-05-25] efa4972 phase 8R-7 A2: hero image srcset + sizes (430ms savings target)
[2026-05-25] 3e19a49 phase 8R-7 A1: Space Grotesk preload + display:optional (LCP target)
[2026-05-25] 93f16a9 phase 8R-6: chunk audit + per-route splitting (706kB -> 268kB delivered, -62%)
[2026-05-25] 53c5d65 phase 8R-5: CmdPalette mount on first Ctrl-K trigger
[2026-05-25] 042dadb phase 8R-4: critical CSS inline + async non-critical via experimental.optimizeCss + critters
[2026-05-25] f58e268 phase 8R-3: atlas-globe priority audit — dropped (below-fold on home, absent on verify/founders)
[2026-05-25] cfd66ae phase 8R-2: HowItWorks / FeatureBento / Manifesto to RSC (via home page RSC conversion)
[2026-05-25] 3972218 phase 8R-1: lazy-load @sentry/nextjs client SDK off initial bundle
[2026-05-25] 350679b Thursday Phase 7: /api/health + admin metrics + webhook duration helper
[2026-05-25] fcd2fb2 Thursday Phase 6: perf optimizations
[2026-05-25] 0f65212 Thursday Phase 5: receipts wall (RSC shell + client island + Supabase Realtime, 0002 migration)
[2026-05-25] 98520ee Thursday Phase 4: Sentry SDK (client+server+edge, replay, global-error, gated test route)
[2026-05-25] 25b41ea polish: pricing order + quote.integrity eyebrow + atlas-globe full figure (3 design canvas reconciliations)
[2026-05-25] b058fd7 Thursday Phase 3: Upstash rate limiting (signup 5/hr, claim 3/hr, api 60/min)
[2026-05-24] d0ef970 Thursday Phase 2: security headers (HSTS, CSP, XFO, XCTO, Referrer, Permissions) via next.config.ts headers()
[2026-05-24] 1531c8b Thursday Phase 1: env staging (Upstash + Sentry creds in /root/MASTERCLAW/.env, mirrored to agentspeech/.env.local — both chmod 600, .env relocked +i)
[2026-05-24] b9dd66f WEDNESDAY_COMPLETE: Tue Supabase data layer + Wed founder claim wizard + tests/contracts
[2026-05-23] f279c7f AgentSpeech.ai pixel port — Sunday foundation + Monday polish
[2026-05-19] 6cd4c2d feat: Bazaar MCP tool — agentspeech_verify_agent (discovery channel 3, read-only)
[2026-05-19] 9424878 plugin stack: audit mcp-builder + skill-creator (already in example-skills bundle; added redundant symlinks)
[2026-05-11] 1ea5ac0 AgentSpeech Day 1 — workspace init, 8/9-plugin power stack, CLAUDE.md doctrine locked
// refreshes every 5 minutes · git is the source of truth