2026-03-13 23:21:55 +00:00
|
|
|
import { Router, type IRouter } from "express";
|
Task #3: MVP API — payment-gated jobs + demo endpoint
OpenAPI spec (lib/api-spec/openapi.yaml)
- Added POST /jobs, GET /jobs/{id}, GET /demo endpoints
- Added schemas: CreateJobRequest, CreateJobResponse, JobStatusResponse,
InvoiceInfo, JobState, DemoResponse, ErrorResponse
- Ran codegen: generated CreateJobBody, GetJobResponse, RunDemoQueryParams etc.
Jobs router (artifacts/api-server/src/routes/jobs.ts)
- POST /jobs: validates body, creates LNbits eval invoice, inserts job +
invoice in a DB transaction, returns { jobId, evalInvoice }
- GET /jobs/:id: fetches job, calls advanceJob() helper, returns state-
appropriate payload (eval/work invoice, reason, result, errorMessage)
- advanceJob() state machine:
- awaiting_eval_payment: checks LNbits, atomically marks paid + advances
state via optimistic WHERE state='awaiting_eval_payment'; runs
AgentService.evaluateRequest, branches to awaiting_work_payment or rejected
- awaiting_work_payment: same pattern for work invoice, runs
AgentService.executeWork, advances to complete
- Any agent/LNbits error transitions job to failed
Demo router (artifacts/api-server/src/routes/demo.ts)
- GET /demo?request=...: in-memory rate limiter (5 req/hour per IP)
- Explicit guard for missing request param (coerce.string() workaround)
- Calls AgentService.executeWork directly, returns { result }
Dev router (artifacts/api-server/src/routes/dev.ts)
- POST /dev/stub/pay/:paymentHash: marks stub invoice paid in-memory
- Only mounted when NODE_ENV !== 'production'
Route index updated to mount all three routers
replit.md: documented full curl flow with all 6 steps, demo endpoint,
and dev stub-pay trigger
End-to-end verified with curl:
- Full flow: create → eval pay → evaluating → work pay → executing → complete
- Error cases: 400 on missing body/param, 404 on unknown job
2026-03-18 15:31:26 +00:00
|
|
|
import healthRouter from "./health.js";
|
|
|
|
|
import jobsRouter from "./jobs.js";
|
Task #5: Lightning-gated node bootstrap (proof-of-concept)
Pay a Lightning invoice → Timmy auto-provisions a Bitcoin full node on DO.
New: lib/db/src/schema/bootstrap-jobs.ts
- bootstrap_jobs table: id, state, amountSats, paymentHash, paymentRequest,
dropletId, nodeIp, tailscaleHostname, lnbitsUrl, sshPrivateKey,
sshKeyDelivered (bool), errorMessage, createdAt, updatedAt
- States: awaiting_payment | provisioning | ready | failed
- Payment data stored inline (no FK to jobs/invoices tables — separate entity)
- db:push applied to create table in Postgres
New: artifacts/api-server/src/lib/provisioner.ts
- ProvisionerService: stubs when DO_API_TOKEN absent, real otherwise
- Stub mode: generates a real RSA 4096-bit SSH keypair via ssh-keygen,
returns RFC 5737 test IP + fake Tailscale hostname after 2s delay
- Real mode: upload SSH public key to DO → generate Tailscale auth key →
create DO droplet with cloud-init user_data → poll for public IP (2 min)
- buildCloudInitScript(): non-interactive bash that installs Docker + Tailscale
+ UFW + Bitcoin Knots via docker-compose; joins Tailscale if authkey provided
- provision() designed as fire-and-forget (void); updates DB to ready/failed
New: artifacts/api-server/src/routes/bootstrap.ts
- POST /api/bootstrap: create job + LNbits invoice, return paymentRequest
- GET /api/bootstrap/:id: poll-driven state machine
* awaiting_payment: checks payment, fires provisioner on confirm
* provisioning: returns progress message
* ready: delivers credentials; SSH private key delivered once then cleared
* failed: returns error message
- Stub mode message includes the exact /dev/stub/pay URL for easy testing
- nextSteps array guides user through post-provision setup
Updated: artifacts/api-server/src/lib/pricing.ts
- Added bootstrapFee field reading BOOTSTRAP_FEE_SATS env var (default 10000)
- calculateBootstrapFeeSats() method
Updated: artifacts/api-server/src/routes/index.ts
- Mounts bootstrapRouter
Updated: replit.md
- Documents all 7 new env vars (DO_API_TOKEN, DO_REGION, DO_SIZE, etc.)
- Full curl-based flow example with annotated response shape
End-to-end verified in stub mode: POST → pay → provisioning → ready (SSH key)
→ second GET clears key and shows sshKeyNote
2026-03-18 18:47:48 +00:00
|
|
|
import bootstrapRouter from "./bootstrap.js";
|
2026-03-18 20:00:24 +00:00
|
|
|
import sessionsRouter from "./sessions.js";
|
Task #3: MVP API — payment-gated jobs + demo endpoint
OpenAPI spec (lib/api-spec/openapi.yaml)
- Added POST /jobs, GET /jobs/{id}, GET /demo endpoints
- Added schemas: CreateJobRequest, CreateJobResponse, JobStatusResponse,
InvoiceInfo, JobState, DemoResponse, ErrorResponse
- Ran codegen: generated CreateJobBody, GetJobResponse, RunDemoQueryParams etc.
Jobs router (artifacts/api-server/src/routes/jobs.ts)
- POST /jobs: validates body, creates LNbits eval invoice, inserts job +
invoice in a DB transaction, returns { jobId, evalInvoice }
- GET /jobs/:id: fetches job, calls advanceJob() helper, returns state-
appropriate payload (eval/work invoice, reason, result, errorMessage)
- advanceJob() state machine:
- awaiting_eval_payment: checks LNbits, atomically marks paid + advances
state via optimistic WHERE state='awaiting_eval_payment'; runs
AgentService.evaluateRequest, branches to awaiting_work_payment or rejected
- awaiting_work_payment: same pattern for work invoice, runs
AgentService.executeWork, advances to complete
- Any agent/LNbits error transitions job to failed
Demo router (artifacts/api-server/src/routes/demo.ts)
- GET /demo?request=...: in-memory rate limiter (5 req/hour per IP)
- Explicit guard for missing request param (coerce.string() workaround)
- Calls AgentService.executeWork directly, returns { result }
Dev router (artifacts/api-server/src/routes/dev.ts)
- POST /dev/stub/pay/:paymentHash: marks stub invoice paid in-memory
- Only mounted when NODE_ENV !== 'production'
Route index updated to mount all three routers
replit.md: documented full curl flow with all 6 steps, demo endpoint,
and dev stub-pay trigger
End-to-end verified with curl:
- Full flow: create → eval pay → evaluating → work pay → executing → complete
- Error cases: 400 on missing body/param, 404 on unknown job
2026-03-18 15:31:26 +00:00
|
|
|
import demoRouter from "./demo.js";
|
|
|
|
|
import devRouter from "./dev.js";
|
2026-03-18 17:43:01 +00:00
|
|
|
import testkitRouter from "./testkit.js";
|
2026-03-20 01:29:50 +00:00
|
|
|
import { lnbitsService } from "../lib/lnbits.js";
|
2026-03-18 18:06:44 +00:00
|
|
|
import uiRouter from "./ui.js";
|
2026-03-18 21:58:41 +00:00
|
|
|
import nodeDiagnosticsRouter from "./node-diagnostics.js";
|
2026-03-18 21:01:13 -04:00
|
|
|
import metricsRouter from "./metrics.js";
|
2026-03-18 22:15:46 -04:00
|
|
|
import worldRouter from "./world.js";
|
2026-03-19 15:59:14 +00:00
|
|
|
import identityRouter from "./identity.js";
|
Task #27: Cost-routing + free-tier gate
## What was built
### DB schema
- `timmy_config` table: key/value store for the generosity pool balance
- `free_tier_grants` table: immutable audit log of every Timmy-absorbed request
- `jobs.free_tier` (boolean) + `jobs.absorbed_sats` (integer) columns
### FreeTierService (`lib/free-tier.ts`)
- Per-tier daily sats budgets (new=0, established=50, trusted=200, elite=1000)
— all env-var overridable
- `decide(pubkey, estimatedSats)` → `{ serve: free|partial|gate, absorbSats, chargeSats }`
— checks pool balance AND identity daily budget atomically
- `credit(paidSats)` — credits POOL_CREDIT_PCT (default 10%) of every paid
work invoice back to the generosity pool
- `recordGrant(pubkey, reqHash, absorbSats)` — DB transaction: deducts pool,
updates identity daily absorption counter, writes audit row
- `poolStatus()` — snapshot for metrics/monitoring
### Route integration
- `POST /api/jobs` (eval → work flow): after eval passes, `freeTierService.decide()`
intercepts. Free → skip invoice, fire work directly. Partial → discounted invoice.
Gate (anonymous/new tier/pool empty) → unchanged full-price flow.
- `POST /api/sessions/:id/request`: after compute, free-tier discount applied to
balance debit. Session balance only reduced by `chargeSats`; absorbed portion
comes from pool.
- Pool credited on every paid work completion (both jobs and session paths).
- Response fields: `free_tier: true`, `absorbed_sats: N` when applicable.
### GET /api/estimate
- Lightweight pre-flight cost estimator; no payment required
- Returns: estimatedSats, btcPriceUsd, tokenEstimate, identity.free_tier decision
(if valid nostr_token provided), pool.balanceSats, pool.dailyBudgets
### Tests
- All 29 existing testkit tests pass (0 failures)
- Anonymous/new-tier users hit gate path correctly (verified manually)
- Pool seeds to 10,000 sats on first boot
## Architecture notes
- Free tier decision happens BEFORE invoice creation for jobs (save user the click)
- Partial grant recorded at invoice creation time (reserves pool capacity proactively)
- Free tier for sessions decided AFTER compute (actual cost known, applied to debit)
- Pool crediting is fire-and-forget (non-blocking)
2026-03-19 16:34:05 +00:00
|
|
|
import estimateRouter from "./estimate.js";
|
task/30: Sovereign Nostr relay infrastructure (strfry)
## Summary
Deploys strfry (C++ Nostr relay) + relay-policy sidecar as a containerised
stack on the VPS, wired to the API server for event-level access control.
## Files created
- `infrastructure/strfry.conf` — strfry config: bind 0.0.0.0:7777, writePolicy
plugin → /usr/local/bin/relay-policy-plugin, maxEventSize 65536,
rejectEphemeral false, db /data/strfry-db
- `infrastructure/relay-policy/plugin.sh` — strfry write-policy plugin (stdin/stdout
bridge). Reads JSON lines from strfry, POSTs to relay-policy HTTP sidecar
(http://relay-policy:3080/decide), writes decision to stdout. Safe fallback:
reject on sidecar timeout/failure
- `infrastructure/relay-policy/index.ts` — Node.js HTTP relay-policy sidecar:
POST /decide receives strfry events, calls API server /api/relay/policy with
Bearer RELAY_POLICY_SECRET, returns strfry decision JSON
- `infrastructure/relay-policy/package.json + tsconfig.json` — TS build config
- `infrastructure/relay-policy/Dockerfile` — multi-stage: builder (tsc) + runtime
- `infrastructure/relay-policy/.gitignore` — excludes node_modules, dist
- `artifacts/api-server/src/routes/relay.ts` — POST /api/relay/policy: internal
route protected by RELAY_POLICY_SECRET Bearer token. Bootstrap state: rejects
all events with "relay not yet open — whitelist pending (Task #37)". Stable
contract — future tasks extend evaluatePolicy() without API shape changes
## Files modified
- `infrastructure/docker-compose.yml` — adds relay-policy + strfry services on
node-net; strfry_data volume (bind-mounted at /data/strfry); relay-policy
healthcheck; strfry depends on relay-policy healthy
- `infrastructure/ops.sh` — adds relay:logs, relay:restart, relay:status commands
- `artifacts/api-server/src/routes/index.ts` — registers relayRouter
## Operator setup required on VPS
mkdir -p /data/strfry && chmod 700 /data/strfry
echo "RELAY_API_URL=https://alexanderwhitestone.com" >> /opt/timmy-node/.env
echo "RELAY_POLICY_SECRET=$(openssl rand -hex 32)" >> /opt/timmy-node/.env
# Also set RELAY_POLICY_SECRET in Replit secrets for API server
## Notes
- TypeScript: 0 errors (API server + relay-policy sidecar both compile clean)
- POST /api/relay/policy smoke test: correct bootstrap reject response
- strfry image: ghcr.io/hoytech/strfry:latest
2026-03-19 20:02:00 +00:00
|
|
|
import relayRouter from "./relay.js";
|
task/31: Relay account whitelist + trust-gated access
## What was built
Full relay access control system: relay_accounts table, RelayAccountService,
trust hook integration, live policy enforcement, admin CRUD API, Timmy seed.
## DB change
`lib/db/src/schema/relay-accounts.ts` — new `relay_accounts` table:
pubkey (PK, FK → nostr_identities.pubkey ON DELETE CASCADE),
access_level ("none"|"read"|"write"), granted_by ("manual"|"auto-tier"),
granted_at, revoked_at (nullable), notes. Pushed via `pnpm run push`.
`lib/db/src/schema/index.ts` — exports relay-accounts.
## RelayAccountService (`artifacts/api-server/src/lib/relay-accounts.ts`)
- getAccess(pubkey) → RelayAccessLevel (none if missing or revoked)
- grant(pubkey, level, reason, grantedBy) — upsert; creates nostr_identity FK
- revoke(pubkey, reason) — sets revokedAt, access_level → none
- syncFromTrustTier(pubkey, tier) — auto-promotes by tier; never downgrades manual grants
- list(opts) — returns all accounts, optionally filtered to active
- Tier→access map: new=none, established/trusted/elite=write (env-overridable)
## Trust hook (`artifacts/api-server/src/lib/trust.ts`)
recordSuccess + recordFailure both call syncFromTrustTier after writing tier.
Failure is caught + logged (non-blocking — trust flow never fails on relay error).
## Policy endpoint (`artifacts/api-server/src/routes/relay.ts`)
evaluatePolicy() now async: queries relay_accounts.getAccess(pubkey).
"write" → accept; "read"/"none"/missing → reject with clear message.
DB error → reject with "policy service error" (safe fail-closed).
## Admin routes (`artifacts/api-server/src/routes/admin-relay.ts`)
ADMIN_SECRET Bearer token auth (localhost-only fallback in dev; error log in prod).
GET /api/admin/relay/accounts — list all accounts
POST /api/admin/relay/accounts/:pk/grant — grant (level + notes body)
POST /api/admin/relay/accounts/:pk/revoke — revoke (reason body)
pubkey validation: must be 64-char lowercase hex.
## Startup seed (`artifacts/api-server/src/index.ts`)
On every startup: grants Timmy's own pubkeyHex "write" access ("manual").
Idempotent upsert — safe across restarts.
## Smoke test results (all pass)
- Timmy pubkey → accept ✓; unknown pubkey → reject ✓
- Admin grant → accept ✓; admin revoke → reject ✓; admin list shows accounts ✓
- TypeScript: 0 errors in API server + lib/db
2026-03-19 20:21:12 +00:00
|
|
|
import adminRelayRouter from "./admin-relay.js";
|
task/32: Event moderation queue + Timmy AI review
## What was built
Full moderation pipeline: relay_event_queue table, strfry inject helper,
ModerationService with Claude haiku review, policy tier routing, 30s poll loop,
admin approve/reject/list endpoints.
## DB schema (`lib/db/src/schema/relay-event-queue.ts`)
relay_event_queue: event_id (PK), pubkey (FK → nostr_identities), kind,
raw_event (text JSON), status (pending/approved/rejected/auto_approved),
reviewed_by (timmy_ai/admin/null), review_reason, created_at, decided_at.
Exported from schema/index.ts. Pushed via pnpm run push.
## strfry HTTP client (`artifacts/api-server/src/lib/strfry.ts`)
injectEvent(rawEventJson) — POST {STRFRY_URL}/import (NDJSON).
STRFRY_URL defaults to "http://strfry:7777" (Docker internal network).
5s timeout; graceful failure in dev when strfry not running; never throws.
## ModerationService (`artifacts/api-server/src/lib/moderation.ts`)
- enqueue(event) — insert pending row; idempotent onConflictDoNothing
- autoReview(eventId) — Claude haiku prompt: approve or flag. On flag, marks
reviewedBy=timmy_ai and leaves pending for admin. On approve, calls decide().
- decide(eventId, status, reason, reviewedBy) — updates DB + calls injectEvent
- processPending(limit=10) — batch poll: auto-review up to limit pending events
- Stub mode: auto-approves all events when Anthropic key absent
## Policy endpoint update (`artifacts/api-server/src/routes/relay.ts`)
Tier routing in evaluatePolicy:
read/none → reject (unchanged)
write + elite tier → injectEvent + accept (elite bypass; shadowReject if inject fails)
write + non-elite → enqueue + shadowReject (held for moderation)
Imports db/nostrIdentities directly for tier check. Both inject and enqueue errors
are fail-closed (reject vs shadowReject respectively).
## Background poll loop (`artifacts/api-server/src/index.ts`)
setInterval every 30s calling moderationService.processPending(10).
Interval configurable via MODERATION_POLL_MS env var.
Errors caught per-event; poll loop never crashes the server.
## Admin queue routes (`artifacts/api-server/src/routes/admin-relay-queue.ts`)
ADMIN_SECRET Bearer auth (same pattern as admin-relay.ts).
GET /api/admin/relay/queue?status=... — list all / by status
POST /api/admin/relay/queue/:eventId/approve — approve + inject into strfry
POST /api/admin/relay/queue/:eventId/reject — reject (no inject)
409 on duplicate decisions. Registered in routes/index.ts.
## Smoke tests (all pass)
Unknown → reject ✓; elite → shadowReject (strfry unavailable in dev) ✓;
non-elite write → shadowReject + pending in queue ✓; admin approve → approved ✓;
moderation poll loop started ✓; TypeScript 0 errors.
2026-03-19 20:35:39 +00:00
|
|
|
import adminRelayQueueRouter from "./admin-relay-queue.js";
|
2026-03-20 02:41:12 +00:00
|
|
|
import geminiRouter from "./gemini.js";
|
2026-03-13 23:21:55 +00:00
|
|
|
|
|
|
|
|
const router: IRouter = Router();
|
|
|
|
|
|
|
|
|
|
router.use(healthRouter);
|
2026-03-18 21:01:13 -04:00
|
|
|
router.use(metricsRouter);
|
Task #3: MVP API — payment-gated jobs + demo endpoint
OpenAPI spec (lib/api-spec/openapi.yaml)
- Added POST /jobs, GET /jobs/{id}, GET /demo endpoints
- Added schemas: CreateJobRequest, CreateJobResponse, JobStatusResponse,
InvoiceInfo, JobState, DemoResponse, ErrorResponse
- Ran codegen: generated CreateJobBody, GetJobResponse, RunDemoQueryParams etc.
Jobs router (artifacts/api-server/src/routes/jobs.ts)
- POST /jobs: validates body, creates LNbits eval invoice, inserts job +
invoice in a DB transaction, returns { jobId, evalInvoice }
- GET /jobs/:id: fetches job, calls advanceJob() helper, returns state-
appropriate payload (eval/work invoice, reason, result, errorMessage)
- advanceJob() state machine:
- awaiting_eval_payment: checks LNbits, atomically marks paid + advances
state via optimistic WHERE state='awaiting_eval_payment'; runs
AgentService.evaluateRequest, branches to awaiting_work_payment or rejected
- awaiting_work_payment: same pattern for work invoice, runs
AgentService.executeWork, advances to complete
- Any agent/LNbits error transitions job to failed
Demo router (artifacts/api-server/src/routes/demo.ts)
- GET /demo?request=...: in-memory rate limiter (5 req/hour per IP)
- Explicit guard for missing request param (coerce.string() workaround)
- Calls AgentService.executeWork directly, returns { result }
Dev router (artifacts/api-server/src/routes/dev.ts)
- POST /dev/stub/pay/:paymentHash: marks stub invoice paid in-memory
- Only mounted when NODE_ENV !== 'production'
Route index updated to mount all three routers
replit.md: documented full curl flow with all 6 steps, demo endpoint,
and dev stub-pay trigger
End-to-end verified with curl:
- Full flow: create → eval pay → evaluating → work pay → executing → complete
- Error cases: 400 on missing body/param, 404 on unknown job
2026-03-18 15:31:26 +00:00
|
|
|
router.use(jobsRouter);
|
Task #27: Cost-routing + free-tier gate
## What was built
### DB schema
- `timmy_config` table: key/value store for the generosity pool balance
- `free_tier_grants` table: immutable audit log of every Timmy-absorbed request
- `jobs.free_tier` (boolean) + `jobs.absorbed_sats` (integer) columns
### FreeTierService (`lib/free-tier.ts`)
- Per-tier daily sats budgets (new=0, established=50, trusted=200, elite=1000)
— all env-var overridable
- `decide(pubkey, estimatedSats)` → `{ serve: free|partial|gate, absorbSats, chargeSats }`
— checks pool balance AND identity daily budget atomically
- `credit(paidSats)` — credits POOL_CREDIT_PCT (default 10%) of every paid
work invoice back to the generosity pool
- `recordGrant(pubkey, reqHash, absorbSats)` — DB transaction: deducts pool,
updates identity daily absorption counter, writes audit row
- `poolStatus()` — snapshot for metrics/monitoring
### Route integration
- `POST /api/jobs` (eval → work flow): after eval passes, `freeTierService.decide()`
intercepts. Free → skip invoice, fire work directly. Partial → discounted invoice.
Gate (anonymous/new tier/pool empty) → unchanged full-price flow.
- `POST /api/sessions/:id/request`: after compute, free-tier discount applied to
balance debit. Session balance only reduced by `chargeSats`; absorbed portion
comes from pool.
- Pool credited on every paid work completion (both jobs and session paths).
- Response fields: `free_tier: true`, `absorbed_sats: N` when applicable.
### GET /api/estimate
- Lightweight pre-flight cost estimator; no payment required
- Returns: estimatedSats, btcPriceUsd, tokenEstimate, identity.free_tier decision
(if valid nostr_token provided), pool.balanceSats, pool.dailyBudgets
### Tests
- All 29 existing testkit tests pass (0 failures)
- Anonymous/new-tier users hit gate path correctly (verified manually)
- Pool seeds to 10,000 sats on first boot
## Architecture notes
- Free tier decision happens BEFORE invoice creation for jobs (save user the click)
- Partial grant recorded at invoice creation time (reserves pool capacity proactively)
- Free tier for sessions decided AFTER compute (actual cost known, applied to debit)
- Pool crediting is fire-and-forget (non-blocking)
2026-03-19 16:34:05 +00:00
|
|
|
router.use(estimateRouter);
|
Task #5: Lightning-gated node bootstrap (proof-of-concept)
Pay a Lightning invoice → Timmy auto-provisions a Bitcoin full node on DO.
New: lib/db/src/schema/bootstrap-jobs.ts
- bootstrap_jobs table: id, state, amountSats, paymentHash, paymentRequest,
dropletId, nodeIp, tailscaleHostname, lnbitsUrl, sshPrivateKey,
sshKeyDelivered (bool), errorMessage, createdAt, updatedAt
- States: awaiting_payment | provisioning | ready | failed
- Payment data stored inline (no FK to jobs/invoices tables — separate entity)
- db:push applied to create table in Postgres
New: artifacts/api-server/src/lib/provisioner.ts
- ProvisionerService: stubs when DO_API_TOKEN absent, real otherwise
- Stub mode: generates a real RSA 4096-bit SSH keypair via ssh-keygen,
returns RFC 5737 test IP + fake Tailscale hostname after 2s delay
- Real mode: upload SSH public key to DO → generate Tailscale auth key →
create DO droplet with cloud-init user_data → poll for public IP (2 min)
- buildCloudInitScript(): non-interactive bash that installs Docker + Tailscale
+ UFW + Bitcoin Knots via docker-compose; joins Tailscale if authkey provided
- provision() designed as fire-and-forget (void); updates DB to ready/failed
New: artifacts/api-server/src/routes/bootstrap.ts
- POST /api/bootstrap: create job + LNbits invoice, return paymentRequest
- GET /api/bootstrap/:id: poll-driven state machine
* awaiting_payment: checks payment, fires provisioner on confirm
* provisioning: returns progress message
* ready: delivers credentials; SSH private key delivered once then cleared
* failed: returns error message
- Stub mode message includes the exact /dev/stub/pay URL for easy testing
- nextSteps array guides user through post-provision setup
Updated: artifacts/api-server/src/lib/pricing.ts
- Added bootstrapFee field reading BOOTSTRAP_FEE_SATS env var (default 10000)
- calculateBootstrapFeeSats() method
Updated: artifacts/api-server/src/routes/index.ts
- Mounts bootstrapRouter
Updated: replit.md
- Documents all 7 new env vars (DO_API_TOKEN, DO_REGION, DO_SIZE, etc.)
- Full curl-based flow example with annotated response shape
End-to-end verified in stub mode: POST → pay → provisioning → ready (SSH key)
→ second GET clears key and shows sshKeyNote
2026-03-18 18:47:48 +00:00
|
|
|
router.use(bootstrapRouter);
|
2026-03-18 20:00:24 +00:00
|
|
|
router.use(sessionsRouter);
|
2026-03-19 15:59:14 +00:00
|
|
|
router.use(identityRouter);
|
task/30: Sovereign Nostr relay infrastructure (strfry)
## Summary
Deploys strfry (C++ Nostr relay) + relay-policy sidecar as a containerised
stack on the VPS, wired to the API server for event-level access control.
## Files created
- `infrastructure/strfry.conf` — strfry config: bind 0.0.0.0:7777, writePolicy
plugin → /usr/local/bin/relay-policy-plugin, maxEventSize 65536,
rejectEphemeral false, db /data/strfry-db
- `infrastructure/relay-policy/plugin.sh` — strfry write-policy plugin (stdin/stdout
bridge). Reads JSON lines from strfry, POSTs to relay-policy HTTP sidecar
(http://relay-policy:3080/decide), writes decision to stdout. Safe fallback:
reject on sidecar timeout/failure
- `infrastructure/relay-policy/index.ts` — Node.js HTTP relay-policy sidecar:
POST /decide receives strfry events, calls API server /api/relay/policy with
Bearer RELAY_POLICY_SECRET, returns strfry decision JSON
- `infrastructure/relay-policy/package.json + tsconfig.json` — TS build config
- `infrastructure/relay-policy/Dockerfile` — multi-stage: builder (tsc) + runtime
- `infrastructure/relay-policy/.gitignore` — excludes node_modules, dist
- `artifacts/api-server/src/routes/relay.ts` — POST /api/relay/policy: internal
route protected by RELAY_POLICY_SECRET Bearer token. Bootstrap state: rejects
all events with "relay not yet open — whitelist pending (Task #37)". Stable
contract — future tasks extend evaluatePolicy() without API shape changes
## Files modified
- `infrastructure/docker-compose.yml` — adds relay-policy + strfry services on
node-net; strfry_data volume (bind-mounted at /data/strfry); relay-policy
healthcheck; strfry depends on relay-policy healthy
- `infrastructure/ops.sh` — adds relay:logs, relay:restart, relay:status commands
- `artifacts/api-server/src/routes/index.ts` — registers relayRouter
## Operator setup required on VPS
mkdir -p /data/strfry && chmod 700 /data/strfry
echo "RELAY_API_URL=https://alexanderwhitestone.com" >> /opt/timmy-node/.env
echo "RELAY_POLICY_SECRET=$(openssl rand -hex 32)" >> /opt/timmy-node/.env
# Also set RELAY_POLICY_SECRET in Replit secrets for API server
## Notes
- TypeScript: 0 errors (API server + relay-policy sidecar both compile clean)
- POST /api/relay/policy smoke test: correct bootstrap reject response
- strfry image: ghcr.io/hoytech/strfry:latest
2026-03-19 20:02:00 +00:00
|
|
|
router.use(relayRouter);
|
task/31: Relay account whitelist + trust-gated access
## What was built
Full relay access control system: relay_accounts table, RelayAccountService,
trust hook integration, live policy enforcement, admin CRUD API, Timmy seed.
## DB change
`lib/db/src/schema/relay-accounts.ts` — new `relay_accounts` table:
pubkey (PK, FK → nostr_identities.pubkey ON DELETE CASCADE),
access_level ("none"|"read"|"write"), granted_by ("manual"|"auto-tier"),
granted_at, revoked_at (nullable), notes. Pushed via `pnpm run push`.
`lib/db/src/schema/index.ts` — exports relay-accounts.
## RelayAccountService (`artifacts/api-server/src/lib/relay-accounts.ts`)
- getAccess(pubkey) → RelayAccessLevel (none if missing or revoked)
- grant(pubkey, level, reason, grantedBy) — upsert; creates nostr_identity FK
- revoke(pubkey, reason) — sets revokedAt, access_level → none
- syncFromTrustTier(pubkey, tier) — auto-promotes by tier; never downgrades manual grants
- list(opts) — returns all accounts, optionally filtered to active
- Tier→access map: new=none, established/trusted/elite=write (env-overridable)
## Trust hook (`artifacts/api-server/src/lib/trust.ts`)
recordSuccess + recordFailure both call syncFromTrustTier after writing tier.
Failure is caught + logged (non-blocking — trust flow never fails on relay error).
## Policy endpoint (`artifacts/api-server/src/routes/relay.ts`)
evaluatePolicy() now async: queries relay_accounts.getAccess(pubkey).
"write" → accept; "read"/"none"/missing → reject with clear message.
DB error → reject with "policy service error" (safe fail-closed).
## Admin routes (`artifacts/api-server/src/routes/admin-relay.ts`)
ADMIN_SECRET Bearer token auth (localhost-only fallback in dev; error log in prod).
GET /api/admin/relay/accounts — list all accounts
POST /api/admin/relay/accounts/:pk/grant — grant (level + notes body)
POST /api/admin/relay/accounts/:pk/revoke — revoke (reason body)
pubkey validation: must be 64-char lowercase hex.
## Startup seed (`artifacts/api-server/src/index.ts`)
On every startup: grants Timmy's own pubkeyHex "write" access ("manual").
Idempotent upsert — safe across restarts.
## Smoke test results (all pass)
- Timmy pubkey → accept ✓; unknown pubkey → reject ✓
- Admin grant → accept ✓; admin revoke → reject ✓; admin list shows accounts ✓
- TypeScript: 0 errors in API server + lib/db
2026-03-19 20:21:12 +00:00
|
|
|
router.use(adminRelayRouter);
|
task/32: Event moderation queue + Timmy AI review
## What was built
Full moderation pipeline: relay_event_queue table, strfry inject helper,
ModerationService with Claude haiku review, policy tier routing, 30s poll loop,
admin approve/reject/list endpoints.
## DB schema (`lib/db/src/schema/relay-event-queue.ts`)
relay_event_queue: event_id (PK), pubkey (FK → nostr_identities), kind,
raw_event (text JSON), status (pending/approved/rejected/auto_approved),
reviewed_by (timmy_ai/admin/null), review_reason, created_at, decided_at.
Exported from schema/index.ts. Pushed via pnpm run push.
## strfry HTTP client (`artifacts/api-server/src/lib/strfry.ts`)
injectEvent(rawEventJson) — POST {STRFRY_URL}/import (NDJSON).
STRFRY_URL defaults to "http://strfry:7777" (Docker internal network).
5s timeout; graceful failure in dev when strfry not running; never throws.
## ModerationService (`artifacts/api-server/src/lib/moderation.ts`)
- enqueue(event) — insert pending row; idempotent onConflictDoNothing
- autoReview(eventId) — Claude haiku prompt: approve or flag. On flag, marks
reviewedBy=timmy_ai and leaves pending for admin. On approve, calls decide().
- decide(eventId, status, reason, reviewedBy) — updates DB + calls injectEvent
- processPending(limit=10) — batch poll: auto-review up to limit pending events
- Stub mode: auto-approves all events when Anthropic key absent
## Policy endpoint update (`artifacts/api-server/src/routes/relay.ts`)
Tier routing in evaluatePolicy:
read/none → reject (unchanged)
write + elite tier → injectEvent + accept (elite bypass; shadowReject if inject fails)
write + non-elite → enqueue + shadowReject (held for moderation)
Imports db/nostrIdentities directly for tier check. Both inject and enqueue errors
are fail-closed (reject vs shadowReject respectively).
## Background poll loop (`artifacts/api-server/src/index.ts`)
setInterval every 30s calling moderationService.processPending(10).
Interval configurable via MODERATION_POLL_MS env var.
Errors caught per-event; poll loop never crashes the server.
## Admin queue routes (`artifacts/api-server/src/routes/admin-relay-queue.ts`)
ADMIN_SECRET Bearer auth (same pattern as admin-relay.ts).
GET /api/admin/relay/queue?status=... — list all / by status
POST /api/admin/relay/queue/:eventId/approve — approve + inject into strfry
POST /api/admin/relay/queue/:eventId/reject — reject (no inject)
409 on duplicate decisions. Registered in routes/index.ts.
## Smoke tests (all pass)
Unknown → reject ✓; elite → shadowReject (strfry unavailable in dev) ✓;
non-elite write → shadowReject + pending in queue ✓; admin approve → approved ✓;
moderation poll loop started ✓; TypeScript 0 errors.
2026-03-19 20:35:39 +00:00
|
|
|
router.use(adminRelayQueueRouter);
|
Task #3: MVP API — payment-gated jobs + demo endpoint
OpenAPI spec (lib/api-spec/openapi.yaml)
- Added POST /jobs, GET /jobs/{id}, GET /demo endpoints
- Added schemas: CreateJobRequest, CreateJobResponse, JobStatusResponse,
InvoiceInfo, JobState, DemoResponse, ErrorResponse
- Ran codegen: generated CreateJobBody, GetJobResponse, RunDemoQueryParams etc.
Jobs router (artifacts/api-server/src/routes/jobs.ts)
- POST /jobs: validates body, creates LNbits eval invoice, inserts job +
invoice in a DB transaction, returns { jobId, evalInvoice }
- GET /jobs/:id: fetches job, calls advanceJob() helper, returns state-
appropriate payload (eval/work invoice, reason, result, errorMessage)
- advanceJob() state machine:
- awaiting_eval_payment: checks LNbits, atomically marks paid + advances
state via optimistic WHERE state='awaiting_eval_payment'; runs
AgentService.evaluateRequest, branches to awaiting_work_payment or rejected
- awaiting_work_payment: same pattern for work invoice, runs
AgentService.executeWork, advances to complete
- Any agent/LNbits error transitions job to failed
Demo router (artifacts/api-server/src/routes/demo.ts)
- GET /demo?request=...: in-memory rate limiter (5 req/hour per IP)
- Explicit guard for missing request param (coerce.string() workaround)
- Calls AgentService.executeWork directly, returns { result }
Dev router (artifacts/api-server/src/routes/dev.ts)
- POST /dev/stub/pay/:paymentHash: marks stub invoice paid in-memory
- Only mounted when NODE_ENV !== 'production'
Route index updated to mount all three routers
replit.md: documented full curl flow with all 6 steps, demo endpoint,
and dev stub-pay trigger
End-to-end verified with curl:
- Full flow: create → eval pay → evaluating → work pay → executing → complete
- Error cases: 400 on missing body/param, 404 on unknown job
2026-03-18 15:31:26 +00:00
|
|
|
router.use(demoRouter);
|
2026-03-20 02:41:12 +00:00
|
|
|
router.use("/gemini", geminiRouter);
|
2026-03-18 17:43:01 +00:00
|
|
|
router.use(testkitRouter);
|
2026-03-18 18:06:44 +00:00
|
|
|
router.use(uiRouter);
|
2026-03-18 21:58:41 +00:00
|
|
|
router.use(nodeDiagnosticsRouter);
|
2026-03-18 22:15:46 -04:00
|
|
|
router.use(worldRouter);
|
Task #3: MVP API — payment-gated jobs + demo endpoint
OpenAPI spec (lib/api-spec/openapi.yaml)
- Added POST /jobs, GET /jobs/{id}, GET /demo endpoints
- Added schemas: CreateJobRequest, CreateJobResponse, JobStatusResponse,
InvoiceInfo, JobState, DemoResponse, ErrorResponse
- Ran codegen: generated CreateJobBody, GetJobResponse, RunDemoQueryParams etc.
Jobs router (artifacts/api-server/src/routes/jobs.ts)
- POST /jobs: validates body, creates LNbits eval invoice, inserts job +
invoice in a DB transaction, returns { jobId, evalInvoice }
- GET /jobs/:id: fetches job, calls advanceJob() helper, returns state-
appropriate payload (eval/work invoice, reason, result, errorMessage)
- advanceJob() state machine:
- awaiting_eval_payment: checks LNbits, atomically marks paid + advances
state via optimistic WHERE state='awaiting_eval_payment'; runs
AgentService.evaluateRequest, branches to awaiting_work_payment or rejected
- awaiting_work_payment: same pattern for work invoice, runs
AgentService.executeWork, advances to complete
- Any agent/LNbits error transitions job to failed
Demo router (artifacts/api-server/src/routes/demo.ts)
- GET /demo?request=...: in-memory rate limiter (5 req/hour per IP)
- Explicit guard for missing request param (coerce.string() workaround)
- Calls AgentService.executeWork directly, returns { result }
Dev router (artifacts/api-server/src/routes/dev.ts)
- POST /dev/stub/pay/:paymentHash: marks stub invoice paid in-memory
- Only mounted when NODE_ENV !== 'production'
Route index updated to mount all three routers
replit.md: documented full curl flow with all 6 steps, demo endpoint,
and dev stub-pay trigger
End-to-end verified with curl:
- Full flow: create → eval pay → evaluating → work pay → executing → complete
- Error cases: 400 on missing body/param, 404 on unknown job
2026-03-18 15:31:26 +00:00
|
|
|
|
2026-03-20 01:29:50 +00:00
|
|
|
// Mount dev routes when NOT in production OR when LNbits is in stub mode.
|
|
|
|
|
// Stub mode means there is no real Lightning backend — payments are simulated
|
|
|
|
|
// in-memory. The testkit relies on POST /dev/stub/pay/:hash to simulate payment
|
|
|
|
|
// confirmation, so we expose it whenever stub mode is active regardless of NODE_ENV.
|
|
|
|
|
// In real production with a live LNbits backend, stubMode is false, so these
|
|
|
|
|
// routes remain unexposed.
|
2026-03-20 02:20:35 +00:00
|
|
|
//
|
|
|
|
|
// OPERATIONAL TRADEOFF: Running the testkit against a live node with LNBITS_STUB=true
|
|
|
|
|
// (e.g. for CI without a real Lightning backend) will expose dev routes in that
|
|
|
|
|
// process. This is intentional — stub mode has no real funds, so the exposure
|
|
|
|
|
// is harmless. Never set LNBITS_STUB=true on a node with a real LNbits backend.
|
2026-03-20 01:29:50 +00:00
|
|
|
if (process.env.NODE_ENV !== "production" || lnbitsService.stubMode) {
|
Task #3: MVP API — payment-gated jobs + demo endpoint
OpenAPI spec (lib/api-spec/openapi.yaml)
- Added POST /jobs, GET /jobs/{id}, GET /demo endpoints
- Added schemas: CreateJobRequest, CreateJobResponse, JobStatusResponse,
InvoiceInfo, JobState, DemoResponse, ErrorResponse
- Ran codegen: generated CreateJobBody, GetJobResponse, RunDemoQueryParams etc.
Jobs router (artifacts/api-server/src/routes/jobs.ts)
- POST /jobs: validates body, creates LNbits eval invoice, inserts job +
invoice in a DB transaction, returns { jobId, evalInvoice }
- GET /jobs/:id: fetches job, calls advanceJob() helper, returns state-
appropriate payload (eval/work invoice, reason, result, errorMessage)
- advanceJob() state machine:
- awaiting_eval_payment: checks LNbits, atomically marks paid + advances
state via optimistic WHERE state='awaiting_eval_payment'; runs
AgentService.evaluateRequest, branches to awaiting_work_payment or rejected
- awaiting_work_payment: same pattern for work invoice, runs
AgentService.executeWork, advances to complete
- Any agent/LNbits error transitions job to failed
Demo router (artifacts/api-server/src/routes/demo.ts)
- GET /demo?request=...: in-memory rate limiter (5 req/hour per IP)
- Explicit guard for missing request param (coerce.string() workaround)
- Calls AgentService.executeWork directly, returns { result }
Dev router (artifacts/api-server/src/routes/dev.ts)
- POST /dev/stub/pay/:paymentHash: marks stub invoice paid in-memory
- Only mounted when NODE_ENV !== 'production'
Route index updated to mount all three routers
replit.md: documented full curl flow with all 6 steps, demo endpoint,
and dev stub-pay trigger
End-to-end verified with curl:
- Full flow: create → eval pay → evaluating → work pay → executing → complete
- Error cases: 400 on missing body/param, 404 on unknown job
2026-03-18 15:31:26 +00:00
|
|
|
router.use(devRouter);
|
|
|
|
|
}
|
2026-03-13 23:21:55 +00:00
|
|
|
|
|
|
|
|
export default router;
|