What is a sovereign AI platform?

A sovereign AI platform runs entirely within your own infrastructure. Daylite combines a no-code AI agent builder with built-in compliance (PII redaction, audit logging, FIPS encryption, SBOM) so regulated enterprises can deploy AI without data leaving their network.

How is Daylite different from Portkey or LiteLLM?

Portkey uses eventual consistency for budget checks, causing overruns under parallel agent load. LiteLLM has P99 latency of 28 seconds at 500 RPS and suffered a supply chain attack in March 2026. Daylite uses atomic pre-flight enforcement with zero race conditions.

How do I integrate Daylite with CrewAI or LangGraph?

Change your base_url to api.daylite.ai/v1 and add an X-Daylite-Customer header. Two lines of code. Works with any OpenAI-compatible SDK.

API Reference

OpenAI-compatible gateway + per-tenant cost attribution + real-time circuit breakers

Quick Start (2 minutes)

1. Get API key at daylite.ai/dashboard (free, no card)

2. Change your base_url to https://api.daylite.ai/v1

3. Add X-Daylite-Customer header to track costs per customer

curl — with cost attribution

curl https://api.daylite.ai/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Daylite-Customer: customer_123" \
  -H "X-Daylite-Feature: chat" \
  -d '{"model":"llama-3.3-70b","messages":[{"role":"user","content":"Hello"}]}'

# Response headers include:
# X-Daylite-Cost: 0.000420
# X-Daylite-Tokens-In: 12
# X-Daylite-Tokens-Out: 847
# X-Daylite-Customer: customer_123

Platform Features

Unique to Daylite

Per-Customer Cost Attribution

Pass customer ID and feature name in request headers. Get exact cost in USD in response headers. No middleware, no proxy — built into every request.

python — track cost per customer

from openai import OpenAI

client = OpenAI(
    base_url="https://api.daylite.ai/v1",
    api_key="YOUR_API_KEY",
)

# The extra_headers parameter passes attribution data
response = client.chat.completions.create(
    model="llama-3.3-70b",
    messages=[{"role": "user", "content": "Summarize this document"}],
    extra_headers={
        "X-Daylite-Customer": "customer_123",
        "X-Daylite-Feature": "document-summary",
    },
)

# Response includes cost in headers
# Access via: response.headers["X-Daylite-Cost"]
print(response.choices[0].message.content)

Real-Time Circuit Breakers

Set hard spend limits per tenant. When exceeded, the request is blocked before it reaches the provider — zero cost. Protects against runaway agent loops that can burn thousands in minutes.

curl — set a budget cap

curl -X POST https://api.daylite.ai/v1/budgets \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "customer_id": "customer_123",
    "monthly_cap_usd": 500.00,
    "action_on_cap": "block",
    "alert_thresholds": [0.5, 0.8, 0.9]
  }'

# When tenant exceeds $500/month:
# → API returns 429 "Budget exceeded" — provider never called
# → Alerts fire at 50%, 80%, 90% via webhook
# → Protects against runaway agent loops

Usage API

Query usage breakdown by customer, feature, or model. Build your own dashboards or export CSV.

curl — get usage per customer

# All customers
curl https://api.daylite.ai/v1/usage \
  -H "Authorization: Bearer YOUR_API_KEY"

# Specific customer
curl "https://api.daylite.ai/v1/usage?customer=customer_123" \
  -H "Authorization: Bearer YOUR_API_KEY"

# Returns: totalCost, totalRequests, byFeature, byModel, budget status

Inference API

Base URL

https://api.daylite.ai/v1

Python (OpenAI SDK)

python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.daylite.ai/v1",
    api_key="YOUR_API_KEY",
)

response = client.chat.completions.create(
    model="llama-3.3-70b",
    messages=[{"role": "user", "content": "What is solar curtailment?"}],
)
print(response.choices[0].message.content)

JavaScript (OpenAI SDK)

typescript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.daylite.ai/v1",
  apiKey: "YOUR_API_KEY",
});

const response = await client.chat.completions.create({
  model: "llama-3.3-70b",
  messages: [{ role: "user", content: "What is solar curtailment?" }],
});
console.log(response.choices[0].message.content);

Streaming

python — streaming

stream = client.chat.completions.create(
    model="llama-3.3-70b",
    messages=[{"role": "user", "content": "Explain inference"}],
    stream=True,
    extra_headers={"X-Daylite-Customer": "customer_123"},
)
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

Models & Pricing

Model	Context	Standard (in/out)	Batch (in/out)
llama-3.3-70b	128K	$0.30 / $0.50	$0.20 / $0.35
llama-3.1-8b	128K	$0.04 / $0.06	$0.03 / $0.04
deepseek-v3	128K	$0.35 / $1.00	$0.25 / $0.70

Authentication

header

Authorization: Bearer dl-your-api-key

Free key at dashboard. 50K requests/month free on Developer tier.

Custom Headers

Header	Direction	Description
X-Daylite-Customer	Request	Your customer ID for cost attribution
X-Daylite-Feature	Request	Feature name for cost breakdown
X-Daylite-Cost	Response	Cost of this request in USD
X-Daylite-Tokens-In	Response	Input tokens consumed
X-Daylite-Tokens-Out	Response	Output tokens generated

Error Codes

400Invalid request — check model name and message format

401Invalid API key

402Insufficient credits — top up in dashboard

429Customer budget exceeded — manage via /v1/budgets

502Backend temporarily unavailable

API Endpoints

Method	Endpoint	Description
POST	/v1/chat/completions	Chat completions (streaming supported)
GET	/v1/models	List available models
GET	/v1/usage	Per-customer usage breakdown
POST	/v1/budgets	Set spend caps per customer
POST	/api/checkout	Purchase credit packs (Stripe)

Try it in the Playground