MYRILUM
MYRILUM Developer Platform

Build autonomous agents on sovereign intelligence

One deterministic API for frontier reasoning, multi-model routing, prompt caching, 50% batch savings, and sandboxed Managed Agents—engineered for production reliability.

MYRILUM Sovereign Routing Plane
99.95% SLA
1. Client & SDK LayerTypeScript · Python · Go · REST
2. Sovereign Policy Gateway & Prompt Cache-90% Cache Read
Zero-retention DLPRegional cell pinningBatch 50% scheduler
Frontier
1M ctx · Deep
Code
512K · Agents
Fast
256K · Instant
3. Sandboxed Managed Agent RuntimesCode Exec · Web Search · MCP

Model lineup & transparent API rates

Switch between real-time streaming rates and asynchronous 50% off Batch pricing.

Batch processing (50% off)

MYRILUM Frontier

Flagship
myrilum-frontier-4

Deepest reasoning, multi-step planning, and long-horizon synthesis for high-stakes workflows.

Input / MTok$10
Output / MTok$30
Cache read$1
1M tokensStandard

MYRILUM Code

Best for Agents
myrilum-code-4

Balanced intelligence and speed tuned for repository-scale code generation, diffs, and tool calling.

Input / MTok$3
Output / MTok$15
Cache read$0.3
512K tokensFast

MYRILUM Fast

Ultra-low latency
myrilum-fast-4

Sub-second first-token speed for high-volume classification, extraction, and real-time copilots.

Input / MTok$0.8
Output / MTok$4
Cache read$0.08
256K tokensInstant

MYRILUM Sovereign Router

Multi-model
myrilum-router-auto

Deterministic policy router dispatching across Frontier, Code, and Fast models with automatic failover.

Input / MTok$2.2
Output / MTok$9.6
Cache read$0.22
Up to 1M tokensAdaptive
Managed Agents & Tools

Ship durable agents without managing container fleets

Delegate long-running tool loops, isolated code execution, and live web retrieval to MYRILUM's managed runtime—with deterministic replay logs and built-in budget caps.

View full API & tool pricing

Managed Agents

$0.04per active agent-minute

Stateful, sandboxed agent runtimes with durable checkpoints, tool retries, and policy guardrails.

Web Search

$8.00per 1,000 searches

Live, citation-grounded web retrieval with domain allowlists and freshness filters built into the API.

Code Execution

$0.05per container hour

Ephemeral, network-isolated Linux sandboxes for Python, Node.js, and shell verification inside tool loops.

Interactive Request Inspector

Switch between TypeScript SDK, cURL, and raw JSON payloads and inspect live response telemetry.

POST https://api.myrilum.ai/v1/messages
import { MyrilumClient } from "@myrilum/sdk";

const client = new MyrilumClient({
  apiKey: process.env.MYRILUM_API_KEY,
  region: "global-sovereign",
});

const response = await client.messages.create({
  model: "myrilum-frontier-4",
  max_tokens: 1024,
  thinking: { type: "enabled", budget_tokens: 2048 },
  messages: [
    {
      role: "user",
      content: "Evaluate failover quorum risk for a 3-region Postgres cluster.",
    },
  ],
});

console.log(response.output_text);
Response Preview
200 OK · Run #1
Latency
640 ms
Input / Output
412 / 186
Cache hit
320 tok
{
  "id": "msg_019b7a4c2f8e",
  "model": "myrilum-frontier-4",
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 412,
    "cache_read_input_tokens": 320,
    "output_tokens": 186
  },
  "content": [
    {
      "type": "text",
      "text": "A 3-region Raft/Patroni quorum tolerates 1 full region outage while preserving RPO=0. To prevent split-brain during WAN partition, pin synchronous_commit=remote_apply and isolate the witness node in an independent availability zone."
    }
  ]
}

Built for your production stack

Integrate MYRILUM natively across cloud providers, observability suites, edge gateways, and data warehouses.

Cloud Infrastructure
AWS PrivateLink

Private VPC peering & regional endpoints

Cloud Infrastructure
Google Cloud

Cross-cloud workload identity federation

Cloud Infrastructure
Microsoft Azure

Entra ID SSO & enterprise networking

Edge Runtime
Cloudflare Workers

Sub-50ms edge gateway & AI caching

Web Framework
Vercel AI SDK

First-class streaming provider adapter

Developer CI/CD
GitHub Actions

Automated PR triage & review workflows

Observability
Datadog

Token telemetry, latency traces & SIEM logs

Data Platform
Snowflake

Governed warehouse SQL & batch pipelines

Product Engineering
Linear

Autonomous issue triage & spec syncing

Billing & FinOps
Stripe

Metered usage webhooks & cost allocation

API Reference & SDKs

Typed TypeScript, Python, and Go SDKs with streaming helpers and tool schemas.

Explore developer tools

Prompt Caching & Batches

Cut token bills up to 90% on long contexts and 50% on asynchronous workloads.

Calculate API costs

Enterprise Security & VPC

Zero-retention routing, regional data residency, and SIEM audit telemetry.

Review security posture
MYRILUM Developer Dispatch

Get model releases, SDK changelogs, and agent engineering notes

Concise monthly technical updates from the MYRILUM Platform team. Unsubscribe anytime.