Stage 04 of 07 in Toolchain
Model Gateway & Router

Sluice

The sovereign model gateway arbitrating all LLM traffic across the ecosystem. A thin LiteLLM proxy on Azure Container Apps that presents one OpenAI-compatible surface to every consumer, while centralising auth, rate limiting, semantic caching, and telemetry in one place. No client needs its own credentials, retry logic, or usage tracking.

Quick Execution · CLI Hook
$ curl -X POST https://sluice.celladoresystems.com/v1/responses
Live (Internal) Verified 2026-08-20
Drop-in OpenAI compatible
existing clients point at Sluice with a base-URL change — no rewrites
Per-workspace virtual keys
RPM, TPM & spend caps enforced at gateway; kill switch doesn't touch consumer code
Redis semantic cache
optional — cuts repeat-prompt cost and latency without changing call sites
Ecosystem Handshakes

How Sluice Connects to the Stack

Celladore is engineered as a unified, composable toolchain where every tool feeds upstream contexts and downstream verification loops without friction.

Gateway OTel Spans → Cost Telemetry

→ Docket

Sluice emits OpenTelemetry spans and Azure OpenAI diagnostic events that Docket ingests for per-project spend attribution.

High-Availability Fallback → Workers

→ Retort

All Retort agent LLM calls route through Sluice, gaining failover, rate limits, and usage attribution without changing a single line of agent code.

Prometheus Metrics → Desktop HUD

→ Deck

Deck reads Sluice's unauthenticated /metrics endpoint directly for gateway health tiles and throughput sparklines in the Dashboard panel.

Architecture Specs

Technical Matrix

Runtime & Engine:
Python 3.12 / LiteLLM Proxy / FastAPI
Deployment Target:
Azure Container Apps (Dynamic Autoscale)
Communication Protocol:
OpenAI REST (/v1/responses, /v1/embeddings, /v1/audio/transcriptions)
Licensing Tier:
Proprietary (Private — Internal Infrastructure)
Upstream Ingestion: Retort Workers, Baton Agents, CodeFlow Engine, Deck HUD, Mystira Story Generator
Downstream Consumers: Docket Telemetry, Azure OpenAI, Application Insights, Langfuse
Venture Origin Bred from Mystira & ConvoLens

Invented from Real Scars

⚡ The Production Incident

Rogue storytelling agent loops racked up thousands in token bills in a single week with zero budget visibility and fatal provider outages disrupting live inference.

🛡️ Failure Mode Prevented

Provider downtime, unmetered API key leakage, and runaway recursive model loops.

Tested in production workloads across Phoenix VC ventures.
Core Capabilities

Architectural Features & Guarantees

Modular design built for high velocity, deterministic reliability, and strict enterprise isolation.

Centralized Auth & Credential Isolation

One shared gateway key replaces Azure OpenAI credentials scattered across every consuming project. A single revocation kills access everywhere instantly.

Rate & Budget Governance

Per-key RPM, TPM, and spend caps enforced at the gateway with a kill switch that doesn't require changes to any consumer code.

Built-in Observability

Prometheus metrics, OpenTelemetry tracing to Application Insights, structured Log Analytics logs, and optional Langfuse LLM observability — without instrumenting callers.

Semantic Response Caching

Optional Redis-backed cache cuts repeat-prompt cost and latency. Identical or semantically near-identical prompts return cached responses transparently.

Per-User Model Selection

FastAPI State Service lets each user select their preferred model from a governed catalog — backed by Azure Cache for Redis in production.

Multi-Environment Deployment

Independent dev, staging, and prod deployments on Azure Container Apps — each with its own Container Apps, Key Vault, and optional Redis instance.

Integration & API Hub

Typed API Contract & Snippets

Live cURL, TypeScript, and Python invocation payloads for Sluice.

DEVELOPER HUB & VERIFIED API CONTRACTS

Unified Developer Interface

Route LLM completions through LiteLLM on Azure Container Apps with per-workspace virtual key metering and budget tagging.

POSThttps://sluice.celladoresystems.com/v1/chat/completionsOpenAI-Compatible
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.CELLADORE_VIRTUAL_KEY || "YOUR_VIRTUAL_KEY",
  baseURL: "https://sluice.celladoresystems.com/v1",
  defaultHeaders: {
    "X-Celladore-Budget-Tag": "project:celladore-web",
  },
});

const response = await client.chat.completions.create({
  model: "claude-3-7-sonnet-latest",
  messages: [
    { role: "system", content: "You are an autonomous engineering agent." },
    { role: "user", content: "Verify PR quality gate requirements and summarize." },
  ],
  temperature: 0.2,
});

console.log(response.choices[0].message.content);

Status Notice: Sluice is live and carrying production traffic across the ecosystem as our internal model gateway. It's infrastructure, not a product with its own signup. Contact us to deploy sovereign gateway infrastructure for your org. Get in touch with the founder .

esc