A.R.C. Service Command Reference

Operator reference for all platform services — ports, health endpoints, make targets, and async interfaces.
Tiers: think | reason | ultra-instinctDev-only profile: observe (think + SigNoz UI, not a tier) Root command: make dev [PROFILE=<tier>]

Infrastructure Services

gateway (`arc-gateway`)

Field	Value
Technology	Traefik v3
Port	`80` (HTTP), `8090` (dashboard)
Health	`http://localhost:8090/ping`
Tiers	think, reason, ultra-instinct

bash

make gateway-up           # start container
make gateway-down         # stop container
make gateway-health       # check health
make gateway-logs         # tail logs
make gateway-nuke         # remove container + image

messaging (`arc-messaging`)

Field	Value
Technology	NATS
Ports	`4222` (client), `8222` (monitoring)
Health	`http://localhost:8222/healthz`
Tiers	think, reason, ultra-instinct

bash

make messaging-up
make messaging-down
make messaging-health
make messaging-logs
make messaging-nuke

streaming (`arc-streaming`)

Field	Value
Technology	Apache Pulsar
Ports	`6650` (broker), `8082` (admin)
Health	`http://localhost:8082/admin/v2/brokers/health`
Tiers	think, reason, ultra-instinct
Note	Cold start ~90 s; `streaming-up` waits up to 120 s

bash

make streaming-up
make streaming-down
make streaming-health
make streaming-logs
make streaming-clean      # interactive confirmation required
make streaming-nuke       # interactive confirmation required

cache (`arc-cache`)

Field	Value
Technology	Redis
Port	`6379`
Health	`docker exec arc-cache redis-cli ping`
Tiers	think, reason, ultra-instinct

bash

make cache-up
make cache-down
make cache-health
make cache-logs
make cache-nuke

persistence (`arc-persistence`)

Field	Value
Technology	PostgreSQL 17 + pgvector
Port	`5432`
Health	`docker exec arc-persistence pg_isready -U arc`
Tiers	think, reason, ultra-instinct

bash

make persistence-up
make persistence-down
make persistence-health
make persistence-logs
make persistence-nuke

storage (`arc-storage`)

Field	Value
Technology	MinIO
Ports	`9000` (API), `9001` (console)
Health	`http://localhost:9000/minio/health/live`
Tiers	reason, ultra-instinct

bash

make storage-up
make storage-down
make storage-health
make storage-logs
make storage-nuke

secrets (`arc-vault`)

Field	Value
Technology	OpenBao
Port	`8200`
Health	`http://localhost:8200/v1/sys/health`
Tiers	reason, ultra-instinct

bash

make vault-up
make vault-down
make vault-health
make vault-logs
make vault-nuke

flags (`arc-flags`)

Field	Value
Technology	Unleash
Port	`4242`
Health	`http://localhost:4242/health`
Tiers	reason, ultra-instinct

bash

make flags-up
make flags-down
make flags-health
make flags-logs
make flags-nuke

otel (`arc-friday`)

Field	Value
Technology	SigNoz + ClickHouse
Ports	`3301` (UI), `8080` (API), `4317` (OTLP gRPC via friday-collector)
Health	`http://localhost:3301/api/v1/health`
Tiers	reason, ultra-instinct
Dev profile	`observe` (think + otel only)

bash

make otel-up
make otel-down
make otel-health
make otel-logs
make otel-nuke

realtime (`arc-realtime`)

Field	Value
Technology	LiveKit Server
Ports	`7880` (HTTP), `7881` (gRPC), `7882` (TCP), `50100–50200` (UDP)
Health	`http://localhost:7880`
Tiers	reason, ultra-instinct
Sidecars	`arc-realtime-ingress` (:7888), `arc-realtime-egress` (:7889)

bash

make realtime-up
make realtime-down
make realtime-health
make realtime-logs
make realtime-nuke

Platform Services

cortex (`arc-cortex`)

Field	Value
Technology	Go
Port	`8801`
Health	`http://localhost:8801/health`
Tiers	think, reason, ultra-instinct (implicit dependency)

bash

make cortex-up
make cortex-down
make cortex-health
make cortex-logs
make cortex-build         # compile Docker image
make cortex-bin           # compile local binary
make cortex-run           # run locally against localhost infra
make cortex-run-dev       # run locally with OTEL enabled
make cortex-test          # go test ./...
make cortex-lint          # golangci-lint
make cortex-nuke

reasoner (`arc-reasoner`)

Field	Value
Technology	Python — FastAPI + LangGraph
Port	`8802`
Health	`http://localhost:8802/health`
Deep health	`http://localhost:8802/health/deep`
Tiers	think, reason, ultra-instinct
Dependencies	persistence, messaging, streaming, friday-collector
Startup timeout	180 s (model download on first cold start)

bash

make reasoner-up
make reasoner-down
make reasoner-health
make reasoner-logs
make reasoner-build
make reasoner-build-fresh
make reasoner-test         # uv run pytest
make reasoner-test-cover   # pytest + coverage
make reasoner-lint         # ruff + mypy
make reasoner-check        # test + lint (mirrors CI)
make reasoner-nuke

HTTP API — OpenAI-compatible endpoints:

Method	Path	Description
`POST`	`/v1/chat/completions`	Chat completion (streaming: `"stream": true`)
`GET`	`/v1/models`	List available models
`POST`	`/v1/responses`	Responses API
`POST`	`/v1/embeddings`	Text embeddings (RAG only)
`POST`	`/v1/files`	Upload file for RAG
`GET`	`/v1/files`	List files
`GET`	`/v1/files/{id}`	Get file metadata
`GET`	`/v1/files/{id}/content`	Stream file content
`DELETE`	`/v1/files/{id}`	Delete file
`POST`	`/v1/vector_stores`	Create vector store
`GET`	`/v1/vector_stores/{id}`	Get vector store
`POST`	`/v1/vector_stores/{id}/files`	Ingest files (async; `?sync=true` available)
`POST`	`/v1/vector_stores/{id}/search`	Search vector store
`GET`	`/health`	Shallow health (always 200)
`GET`	`/health/deep`	Deep health (checks postgres, nats, minio)

Async port interfaces (services/reasoner/src/reasoner/interfaces.py):

python

class LLMProviderPort(Protocol):
    def create_llm(self) -> BaseChatModel: ...
    def supports_system_role(self) -> bool: ...
    def provider_name(self) -> str: ...

class ChatCompletionPort(Protocol):
    async def complete(self, req: ChatCompletionRequest) -> ChatCompletionResponse: ...

class StreamingPort(Protocol):
    def stream(self, req: ChatCompletionRequest) -> AsyncIterator[ChatCompletionChunk]: ...

class ModelRegistryPort(Protocol):
    def list_models(self) -> ModelList: ...
    def model_exists(self, model_id: str) -> bool: ...

class ResponsesPort(Protocol):
    async def respond(self, req: ResponsesRequest) -> ResponsesResponse: ...

class OpenAINATSPort(Protocol):
    async def connect(self) -> None: ...
    async def subscribe(self) -> None: ...
    def is_connected(self) -> bool: ...
    async def close(self) -> None: ...

voice (`arc-voice-agent`)

Field	Value
Technology	Python — FastAPI + LiveKit Agents
Port	`8803`
Health	`http://localhost:8803/health`
Deep health	`http://localhost:8803/health/deep`
Tiers	reason, ultra-instinct
Dependencies	messaging, streaming, realtime, friday-collector, cache
Startup timeout	60 s

bash

make voice-up
make voice-down
make voice-health
make voice-logs
make voice-build
make voice-build-fresh
make voice-test            # uv run pytest
make voice-test-cover      # pytest + coverage
make voice-lint            # ruff + mypy
make voice-check           # test + lint (mirrors CI)
make voice-nuke

HTTP API:

Method	Path	Description
`POST`	`/v1/audio/transcriptions`	Transcribe audio file (STT) — multipart `file` field
`POST`	`/v1/audio/speech`	Synthesize text to WAV (TTS) — JSON body
`GET`	`/health`	Shallow health (always 200)
`GET`	`/health/deep`	Deep health (checks arc-realtime + arc-messaging)

Async port interfaces (services/voice/src/voice/interfaces.py):

python

class STTPort(Protocol):
    async def transcribe(
        self, audio_bytes: bytes, language: str | None = None
    ) -> TranscriptResult: ...
    # TranscriptResult: text, language, duration_secs

class TTSPort(Protocol):
    async def synthesize(self, text: str, voice: str) -> SynthesisResult: ...
    # SynthesisResult: wav_bytes, sample_rate, duration_secs

class LLMBridgePort(Protocol):
    async def reason(
        self, transcript: str, session_id: str, correlation_id: str
    ) -> str: ...
    # Raises: BridgeError(is_timeout, error_type)

Concrete adapters:

Port	Adapter	Implementation
`STTPort`	`WhisperSTTAdapter`	faster-whisper, runs in thread-pool executor
`TTSPort`	`PiperTTSAdapter`	piper binary subprocess, runs in thread-pool executor
`LLMBridgePort`	`NATSBridge`	NATS request-reply to `arc-reasoner`

Global Orchestration

bash

# Start a tier (all services in profile, dependency order)
make dev                         # think profile (default)
make dev PROFILE=observe         # think + SigNoz (dev only)
make dev PROFILE=reason          # full dev stack
make dev PROFILE=ultra-instinct  # everything

# Control
make dev-down          # stop all profile services
make dev-health        # check health of all profile services
make dev-logs          # tail logs from all profile services
make dev-status        # container status table
make dev-images        # check / pull images for profile

# Destructive (confirm before running)
make dev-clean         # remove containers + volumes
make dev-nuke          # remove containers + volumes + images

# Utilities
make help              # list all targets
make scrub             # clean all build caches (Python, Go, Node, Make)
make dev-regen         # force regenerate .make/ profile + registry metadata
make otel-up           # OTEL collector + SigNoz stack (alias for otel targets)

A.R.C. Service Command Reference

Infrastructure Services

gateway (`arc-gateway`)

messaging (`arc-messaging`)

streaming (`arc-streaming`)

cache (`arc-cache`)

persistence (`arc-persistence`)

storage (`arc-storage`)

secrets (`arc-vault`)

flags (`arc-flags`)

otel (`arc-friday`)

realtime (`arc-realtime`)

Platform Services

cortex (`arc-cortex`)

reasoner (`arc-reasoner`)

voice (`arc-voice-agent`)

Global Orchestration

Contributors

Changelog

A.R.C. Service Command Reference ​

Infrastructure Services ​

gateway (arc-gateway) ​

messaging (arc-messaging) ​

streaming (arc-streaming) ​

cache (arc-cache) ​

persistence (arc-persistence) ​

storage (arc-storage) ​

secrets (arc-vault) ​

flags (arc-flags) ​

otel (arc-friday) ​

realtime (arc-realtime) ​

Platform Services ​

cortex (arc-cortex) ​

reasoner (arc-reasoner) ​

voice (arc-voice-agent) ​

Global Orchestration ​

Contributors

Changelog

A.R.C. Service Command Reference

Infrastructure Services

gateway (`arc-gateway`)

messaging (`arc-messaging`)

streaming (`arc-streaming`)

cache (`arc-cache`)

persistence (`arc-persistence`)

storage (`arc-storage`)

secrets (`arc-vault`)

flags (`arc-flags`)

otel (`arc-friday`)

realtime (`arc-realtime`)

Platform Services

cortex (`arc-cortex`)

reasoner (`arc-reasoner`)

voice (`arc-voice-agent`)

Global Orchestration