These were live on the box but never committed. Chat-session replicas of
the deepseek-flash suppliers (OpenRouter deepseek-v4-flash, llm.chat.v1,
ticket settle mode); secrets stay in the gitignored supplier/.env files.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two structural fixes for demo concurrency (3 single-slot suppliers thrashed
under Open WebUI's ~3 conversations per message):
Multi-slot suppliers (MAX_CHAT_SESSIONS, default 1):
- SupplierState is now capacity-aware: a Set of active escrow refs with
free-until-full status semantics, ref-scoped release(ref) (also fixes the
latent bug where handler catch-alls released another request's slot), and
a walletMutex serializing every wallet-spending chain op — Claim, Submit,
and wallet-health consolidation queue instead of racing the single
working/collateral UTxO. /status gains active_sessions/max_sessions.
Entry-ticket settle mode (CHAT_SETTLE_MODE=ticket, default full, both sides):
- The buyer's escrow post at session open is the ONLY chain op. The supplier
verifies the Open escrow as an entry ticket but never Claims/Submits
(/v1/chat/start → {status:"ticket"}; /v1/chat/end → receipt-less
{status:"ended"}); no hard-cap timer, so sessions may outlive deliver_by.
Used tickets can't reopen sessions (records retained; 409 session_exists).
- Buyer SDK: startChat/endChat are settle-mode aware, driven by the
supplier's response — mixed-mode rollout windows settle correctly.
- Gateway: ticket closes bill cost 0; the sweeper Reclaims Open tickets
after deliver_by WITHOUT closing live sessions (the conversation and its
affinity mirror survive their reclaimed ticket), gates reclaims on
deliver_by, and keeps an in-process settled-refs memory so terminal
indexer rows aren't retried every sweep forever.
Deployment (kimi-chat trio + gateway): ticket mode, 8 slots per supplier,
idle policy gateway 10 min / supplier 15 min (fixes the inversion where the
supplier idle-closed before the gateway janitor).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
MVP demo: one operator-funded API key anyone can plug into an IDE/harness,
running real escrow-per-session marketplace conversations (llm.chat.v1)
behind the standard OpenAI /chat/completions wire shape.
- shared: ChatMessage gains tool role + tool_calls; normalizeChatMessage/
chatMessagesEquivalent keep supplier transcript and gateway mirror
hash-identical for receipts (field presence is part of the identity)
- supplier: /v1/chat/message accepts {messages,tools,tool_choice} alongside
legacy {content}; callOpenAiStream forwards tools, accumulates fragmented
delta.tool_calls, allows content-less tool turns; done frame carries the
assistant message verbatim + finish_reason + usage
- gateway: demo flag on api_keys (additive migration), create-demo-key CLI,
withdraw 403 + hidden deposit address for demo keys, per-IP demo rate
limit (demo keys skip the shared per-key limit), models filtered to chat
capability
- gateway executor (demoChat.ts): transcript-prefix affinity maps stateless
calls onto open sessions, forwards only unseen deltas, transparently
reopens on supplier idle-close; never falls back to one-shot (its settle
path assumes <=1 in-flight escrow per wallet)
- mutex scope: key mutex held only for chain work (open/close); off-chain
turns serialize on per-session locks so replicas serve users concurrently
- sweeper: chat-aware reclaim guard (35min vs 11min) and abandoned-session
cleanup; terminal state confirmed via on-chain queryUtxo (indexer keeps
lineage rows at their old state forever)
- deploy: kimi-chat replicas 2+3 compose files; chat idle timeout 2.5min
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Add 3 single-slot suppliers serving deepseek/deepseek-v4-flash via OpenRouter,
increasing concurrent capacity for the independent buyer machines. Same shape
as the kimi replicas: capability llm.text.generate.v1, price 200000 lovelace,
max_output_tokens=262144, max_processing_ms=3600000, OPENAI_TIMEOUT_MS=3600000,
sharing the one OpenRouter key (accepted concurrency ceiling; no 429/backoff).
- deploy/mainnet/docker-compose.supplier-deepseek-flash-{1,2,3}.yml (new)
- buyer/scripts/monitor-wallets.ts: register the 3 replica wallets
Per-instance wallets, .env files, and on-chain adverts are provisioned on the
mainnet host (gitignored); the 3 containers are live and Active on-chain.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add 3 parallel-capacity replicas of supplier-kimi (kimi-2/3/4) so the
marketplace can serve more concurrent jobs from the independent buyer
machines. Each is a single-slot worker on the same capability
(llm.text.generate.v1) and model (moonshotai/kimi-k2.6) at the same 0.2 AP3X
flat price, sharing the one OpenRouter key (accepted concurrency ceiling).
Also bring the original supplier-kimi up to the same config: raise its advert
caps (re-posted on-chain) and OPENAI_TIMEOUT_MS to match. All four kimi
suppliers now advertise max_output_tokens=262144, max_processing_ms=3600000
(OpenRouter's ceiling for kimi-k2.6), with OPENAI_TIMEOUT_MS=3600000 so the
upstream call isn't cut short before the advertised deadline.
- deploy/mainnet/docker-compose.supplier-kimi-{2,3,4}.yml (new)
- deploy/mainnet/docker-compose.supplier-kimi.yml: OPENAI_TIMEOUT_MS 300000 -> 3600000
- buyer/scripts/monitor-wallets.ts: register the 3 replica wallets
Per-instance wallets, .env files, and on-chain adverts are provisioned on the
mainnet host (gitignored); all four containers are live and Active on-chain.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add an "API Keys" page to the buyer SPA (marketplace.vector.apexfusion.org)
that mints a custodial gateway API key by POSTing directly to the gateway's
public /signup, then shows the key once + deposit address + a base_url/curl
usage snippet. Page lives behind the SPA's existing operator login; per-user
registration is a later feature.
Gateway:
- GATEWAY_CORS_ORIGINS env -> corsOrigins allowlist; corsMiddleware echoes
Access-Control-* only for allowlisted Origins, answers OPTIONS preflight
204, no credentials. Empty allowlist = CORS off (default safe).
Buyer:
- GATEWAY_PUBLIC_URL env -> injected into __BUYER_BOOT__.gatewayUrl (only in
the authenticated boot block). SPA resolveGatewayUrl() falls back to
deriving "api." + location.host so the post-login-no-reload case (boot
block absent) still works.
- New ApiKeys page + /api-keys route + nav link.
Deploy (mainnet compose): gateway GATEWAY_CORS_ORIGINS and buyer
GATEWAY_PUBLIC_URL set to the marketplace/api hosts. NOTE: both images must be
REBUILT (gateway runs tsx from copied source; buyer SPA is vite-built at image
build) — an env-only recreate would silently ship the old code.
Typecheck + vite build clean; 205 tests pass. Adversarial multi-agent review
(4 dimensions, per-finding verification) surfaced only the rebuild-required
deploy note above; no code/security defects.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Without VECTOR_ZERO_TIME_MS the live tx builder defaults to the testnet genesis,
placing PostEscrow validity bounds ~51 days in the future (Ogmios 3118). Escrow
spends (Accept/Reclaim) also need ESCROW_REF_UTXO. Mirror buyer/supplier mainnet env.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
New multi-tenant gateway package: external developers point any OpenAI SDK at
the marketplace. Each API key is its own custodial buyer wallet (AES-256-GCM
sealed under GATEWAY_MASTER_KEY); every request posts an on-chain escrow from
that user's wallet.
- POST /openai/v1/chat/completions (one escrow/call, llm.text.generate.v1,
buffered + pseudo-stream), chat sessions (one escrow/session, llm.chat.v1,
real SSE), GET /openai/v1/models, /signup, /account, /account/withdraw.
- Settles by resolving the Submitted ref via the indexer + awaited Accept;
state-aware reclaim sweeper + per-wallet auto-consolidation for collateral.
- Reuses the buyer Marketplace SDK via new buyer exports ./sdk; new shared
buildWithdrawTx (custodial exit). Runs via tsx; Dockerfile + mainnet compose.
Design in docs/gateway.md. Typecheck passes for all packages; 20 vitest tests
pass. Not yet exercised against a live chain.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A one-shot monitor for the vector-marketplace box: checks each wallet in
wallets.json against a threshold (default 10 APEX) via Ogmios
(ReadOnlyOgmiosProvider) and posts a combined Slack message when any wallet is
below it. Re-reminds every 6h while a wallet stays low and posts a recovery
message when it climbs back above, deduped through a persisted state file.
Runs hourly under host cron via `docker compose run --rm` — no long-running
server, no ports/healthcheck. Reuses the buyer/supplier tsx Docker pattern and
the existing balance-query client; addresses come from a static, gitignored
wallets.json so the monitor never reads a private key.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The mainnet buyer compose never set TTS_PIPER_BASE_URL (only testnet did),
so the buyer booted with an empty value and POST /v1/synth-speech — the
no-escrow demo proxy behind the "PiperTTS demo" tile — returned 503
"booted without TTS_PIPER_BASE_URL". Point it at the private piper backend
over internal docker DNS (buyer is already on marketplace-mainnet-net).
Verified: authenticated POST /v1/synth-speech returns 200 audio/mpeg.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds a conversational supplier capability and a free demo, both streaming.
New paid `llm.chat.v1` type: one escrow per conversation (flat fee). "Start
chat" posts the escrow with a session-init placeholder prompt_hash and the
supplier Claims (reserving its single slot); turns then stream off-chain via
SSE with zero chain interaction; "End chat" has the supplier Submit a receipt
over the full transcript and the buyer Accept — the moment the user is charged.
The on-chain validator is content-agnostic (pins but never verifies the
hashes), so this reuses the existing contract unchanged. A supplier-side idle
watchdog auto-ends abandoned sessions before deliver_by.
Free Kimi K2.6 chat demo: a synthetic "demo:kimi-chat" tile that uses the same
chat UI but proxies straight to OpenRouter (no escrow, no payment), mirroring
the existing Piper demo.
Shared:
- postChatEscrow.ts + chatSessionPromptHash() (session-init prompt_hash helper
shared by the buyer builder and the supplier start gate).
Supplier:
- CAPABILITY_KIND="chat-session" (+ CHAT_IDLE_TIMEOUT_MS); callOpenAiStream()
streaming client; ChatSessionStore + endChatSession(); /v1/chat/{start,
message,end} handlers with idle + hard-cap watchdog timers.
Buyer:
- Marketplace.startChat()/endChat() with relaxed (session-init) receipt
verification + crypto verifyReceipt; /v1/chat/{start,message,end} routes
(SSE passthrough) and a free /v1/chat-demo/message OpenRouter proxy;
OPENROUTER_API_KEY/OPENROUTER_BASE_URL config; ChatForm.tsx + Dashboard
KIMI_DEMO_SUPPLIER tile and llm.chat.v1 dispatch.
Deploy:
- docker-compose.supplier-kimi-chat.yml + supplier/.env.kimi-chat.example.
Typecheck passes across all workspaces; full vitest suite green except 15
pre-existing time-bomb/RED failures (tx-claim/accept/post-advert live-CBOR
fixtures with hardcoded past deadlines + the M0-C golden-datum test), confirmed
identical on a clean tree.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The summarization runs server-side, but the SPA held the job only in
component state and jobs lived only in memory — so navigating away from
the Book Summarizer reset the UI (losing the handle to a still-running
job) and past summaries vanished on restart.
- PdfJobDb: SQLite store (its own file under ARCHIVE_DIR) persisting
every job on each progress update.
- JobStore: createdAt, persist-on-update, viewOf()/list() that serve
live jobs first then fall back to the durable record; non-live jobs
whose worker died surface as a new "interrupted" status.
- routes: GET /v1/pdf-jobs (newest-first list); GET /:id + summary.md
serve past jobs via viewOf; /:id/events sends a single terminal frame
for a finished/past job (no hang) and live frames otherwise.
- SPA: RecentJobs list (open any job → reconnect to live progress or
show result), localStorage resume of the active job on return, and a
"back to list (keeps running)" affordance.
- mainnet buyer compose: persistent named volume + ARCHIVE_DIR so the
archive + pdf-jobs.db survive container recreation/redeploys.
Tests: persistence-across-restart, interrupted detection, list, and
past-job serving (incl. terminal SSE). Full buyer suite green (185).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fourth mainnet supplier, branded by model rather than provider so future
OpenRouter-backed suppliers can claim their own compose files without
naming collisions. LLM_BACKEND=openai pointed at https://openrouter.ai/api
with a Bearer-token API key in supplier/.env.kimi. Traefik host
mp-suppliers-kimi.vector.apexfusion.org. No code changes needed — the
OPENAI_API_KEY plumbing landed in 1466442.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds an optional OPENAI_API_KEY env var that supplier/src/openai.ts forwards as
`Authorization: Bearer <key>`. Empty/unset preserves the ChatMock behaviour
(no auth header). Unlocks pointing the openai backend at DeepSeek directly
(https://api.deepseek.com) — no local proxy needed.
New compose: deploy/mainnet/docker-compose.supplier-deepseek.yml (third mainnet
supplier, model deepseek-v4-pro, mp-suppliers-deepseek.vector.apexfusion.org).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- jobRunner.ts: console.warn at each jobs.fail site (chat + tts paths, 6 sites). Three mainnet GPT-supplier 502s on 2026-05-19 left no trace beyond Traefik — adding a one-line log restores grep-ability.
- docker-compose.supplier-gpt.yml: OPENAI_TIMEOUT_MS 180000→300000. The 3 observed 502s were all exactly 90 polls × 2s, i.e. the supplier's openai_timeout firing.
- docker-compose.chatmock.yml: serve --reasoning-effort low --reasoning-summary none. Default 'medium'/'auto' pushed GPT-5.4 think-time over the prior 180s budget.
Note: on-chain advert ADVERT_MAX_PROCESSING_MS still pins 180000 until a re-advert tx is posted — buyers compare against the on-chain value, not the runtime env.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Mirrors deploy/testnet/docker-compose.supplier-tts.yml and the simulation
PiperTTS stack, co-located on vector-marketplace as a private internal
service on marketplace-mainnet-net (no Traefik labels, no public DNS for
Piper itself). Supplier reaches Piper at http://marketplace-mainnet-tts-piper:8000.
Voice: en_US-lessac-high. New supplier wallet + on-chain advert are
published out-of-band by the operator per the runbook header.
Files:
- deploy/mainnet/docker-compose.tts-piper.yml openedai-speech-min, private
- deploy/mainnet/docker-compose.supplier-tts.yml CAPABILITY_KIND=tts supplier
- deploy/mainnet/install-voice-extras.sh voice download + yaml mapping
- deploy/mainnet/voice-mappings.py idempotent yaml editor
- supplier/.env.tts.example operator env template
- .gitignore whitelist .env.*.example
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Restore the canonical mainnet docker-compose.supplier.yml to mirror
testnet's ollama-backed supplier (LLM_BACKEND=ollama, OLLAMA_URL points
at marketplace-mainnet-ollama, Traefik routes mp-suppliers.vector.apex-
fusion.org). The ChatGPT/Codex variant introduced in 965fabd moves to
docker-compose.supplier-gpt.yml with project name + container name
suffixed -gpt so both can coexist on the same host.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The storagetime/chatmock image has neither wget nor curl, so the
previous wget-based healthcheck always failed with "executable file
not found" and the container sat in "starting" forever. The image
does carry a Python interpreter (chatmock is a Flask app), so swap
to a 1-liner urllib probe of /health. /health returns 200 even
before OAuth login, so a red healthcheck now means the Flask server
itself is down.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
storagetime/chatmock on Docker Hub publishes only {latest, 1.36, v1.36}
(all the same digest sha256:757c65c8b28941...). The :v1.35 pin from
the initial commit resolves to "not found". The compose's own comment
forbids :latest, so bump to :v1.36.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a Codex/ChatGPT backend path for the supplier alongside the
existing Ollama one. The mainnet supplier is repointed at a localhost
ChatMock proxy (RayBytes/ChatMock, pinned :v1.35) that authenticates
against a ChatGPT Pro subscription via OAuth.
- supplier/src/openai.ts: HTTP client for OpenAI-compatible
/v1/chat/completions. Same return shape as ollama.callOllama so
jobRunner can branch cleanly.
- supplier/src/jobRunner.ts: branch on config.llmBackend.
- supplier/src/config.ts: LLM_BACKEND ("ollama" default | "openai"),
OPENAI_BASE_URL, OPENAI_TIMEOUT_MS (default 180_000 — GPT first-token
latency).
- supplier/src/cli/gen-keypair.ts: standalone Ed25519 keypair generator
+ bech32 address derivation. Used to bootstrap a fresh supplier
wallet without round-tripping through tx:post-advert --dry-run.
- deploy/mainnet/docker-compose.chatmock.yml: long-running chatmock
serve + one-shot --profile login flow.
- deploy/mainnet/docker-compose.supplier.yml: rewritten for openai
backend; Traefik routes at mp-suppliers-gpt.vector.apexfusion.org so
the unsuffixed slot stays free for a future supplier with a
different backend.
- tests/unit/supplier-openai.test.ts: unit coverage for callOpenAi.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The mainnet host (vector-marketplace, 91.98.147.172) has no local
apex-dashboard-mainnet stack — the buyer + indexer already reach the
chain via the public HTTPS endpoint
https://ogmios.vector.mainnet.apexfusion.org. Pin the supplier to the
same pattern: drop the ws://ogmios:1337 env override (let supplier/.env
supply OGMIOS_URL) and drop the apex-net binding so compose-up doesn't
fail on a missing external network.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Drops the "DEFERRED — Do NOT compose up" banner from
docker-compose.ollama.yml and docker-compose.supplier.yml (mirrors
cd0edd7 for buyer.yml + indexer.yml) and adds the missing LIVE_CHAIN=1
inline override that the testnet supplier compose has at line 45 —
without it the mainnet supplier silently falls back to
ReadOnlyOgmiosProvider and Claim/Submit never broadcast.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
buyer and indexer mainnet stacks have been running on vector-marketplace
(marketplace-mainnet-buyer + marketplace-mainnet-indexer healthy, 4+ days
uptime for indexer, buyer redeployed today with the auth gate). The
"Wired but not yet exercised" header contradicts reality and would spook
the next operator who reads it.
Note in the new headers points at marketplace.vector.apexfusion.org and
mp-indexer.vector.apexfusion.org so the connection between the compose
file and the live URL is explicit. The genuinely-still-open mainnet-
safety follow-up (ARCHITECTURE.md §9 #5: env-gate / wallet allowlist /
MAINNET=1 flag) is kept as a one-line reference in the buyer header.
supplier and ollama mainnet composes are intentionally left alone — I
don't have evidence of those being deployed via these compose files.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Removing the hardcoded OGMIOS_URL=ws://ogmios:1337 (apex-dashboard-mainnet
internal hostname) lets indexer/.env supply the real value — hosted WSS for
deployments that don't run a co-located ogmios container.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The legacy dashboard_default name leaked from apex-dashboard on the
testnet host. On a fresh mainnet box with no apex-dashboard, traefik-net
is more accurate. Mainnet-only — testnet compose files keep
dashboard_default since they depend on the existing apex-dashboard host.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
shared: VECTOR_ZERO_TIME now reads VECTOR_ZERO_TIME_MS from env, defaults
to testnet genesis (1_752_057_484_000 ms). Mainnet deploys set
VECTOR_ZERO_TIME_MS=1756485600000. Existing testnet deployments unaffected
since env unset → testnet default. Patched in both lucidContext.ts (lucid
slot config override) and liveCbor.ts buildLiveTxForSubmit (submitted_at
slot-alignment).
deploy/mainnet: buyer host changed mp.vector.* → marketplace.vector.* per
operator-owned DNS. Indexer compose drops apex-net network (we point at
the hosted mainnet ogmios over public WSS, not via apex-dashboard-mainnet
which is not deployed on this box).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
