Billama 3.0 System Architecture & Developer Guide

// 1. Installation
pnpm add @billama/sdk

// 2. Client Initialization & Streaming Chat
import { Billama } from "@billama/sdk";

const billama = new Billama({
  apiKey: process.env.BILLAMA_API_KEY,
});

const stream = await billama.chat.stream({
  model: "llama3.3:70b",
  messages: [{ role: "user", content: "Optimize our data pipeline." }],
});

for await (const chunk of stream) {
  process.stdout.write(chunk.content);
}

// Access diagnostic routing telemetry:
console.log({
  tokens: stream.usage?.totalTokens,
  latencyMs: stream.latencyMs,
  costUsd: stream.costUsd,
  routerTarget: stream.routeTarget
});

// 3. User Wallet Balance Check & Subscriptions
const balance = await billama.billing.getBalance();
console.log("Credits remaining: $", balance.balanceUsd);

const session = await billama.billing.createCheckoutSession({
  planId: "pro_monthly",
  customerEmail: "user@domain.com",
  successUrl: "https://app.domain.com/settings/billing?success=true",
  cancelUrl: "https://app.domain.com/pricing",
});

from openai import OpenAI

client = OpenAI(
    base_url="https://api.billama.net/v1",
    api_key="blm_live_your_key_here"
)

response = client.chat.completions.create(
    model="llama3.3:70b",
    messages=[
        {"role": "system", "content": "You are a specialized AI assistant."},
        {"role": "user", "content": "Generate a high-concurrency token router."}
    ],
    temperature=0.6,
    stream=True
)

for chunk in response:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

# 1. Chat completions
curl https://api.billama.net/api/chat \
  -H "Authorization: Bearer blm_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-r1:70b",
    "messages": [{"role": "user", "content": "Optimize PostgreSQL connection pooling."}],
    "stream": true
  }'

# 2. Text generation
curl https://api.billama.net/api/generate \
  -H "Authorization: Bearer blm_live_your_key_here" \
  -d '{"model": "llama3.2:8b", "prompt": "Summarize the Billama whitepaper."}'

# 3. Model list
curl https://api.billama.net/api/tags \
  -H "Authorization: Bearer blm_live_your_key_here"

# Step 1: Install daemon via 1-liner
curl -sSL https://billama.net/install.sh | sh

# Step 2: Start worker with your node token
billama-node start \
  --server=https://api.billama.net \
  --token=blm_node_your_node_token \
  --max-vram=22GB \
  --zero-idle=true

# Supported CLI flags:
# --server     URL of Billama Gateway
# --token      Node authentication secret (from dashboard)
# --max-vram   VRAM cap to reserve for Billama workloads
# --zero-idle  Enable automatic cascade to Tier 2/3 jobs when idle
# --port       Local Ollama port (default: 11434)

# Submit 3D rendering batch job via REST
curl -X POST https://api.billama.net/v1/jobs/submit \
  -H "Authorization: Bearer blm_live_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "blender_render",
    "scene_url": "https://assets.billama.net/scenes/demo.blend",
    "resolution": [3840, 2160],
    "samples": 512,
    "tile_grid": [4, 4]
  }'

# Environment Variables for LDAP Integration:
LDAP_URL="ldap://freeipa.freeipa.svc.cluster.local:389"
LDAP_BIND_DN="uid=admin,cn=users,cn=accounts,dc=billama,dc=local"
LDAP_BIND_PASSWORD="AdminPassword123!"
LDAP_SEARCH_BASE="dc=billama,dc=local"
LDAP_USER_SEARCH_FILTER="(&(objectClass=person)(|(uid={{username}})(mail={{username}})))"
LDAP_GROUP_SEARCH_FILTER="(member={{dn}})"

helm repo add billama https://charts.billama.net
helm upgrade --install billama billama/billama \
  --namespace billama \
  --create-namespace \
  --values values-production.yaml