Analytics Headers

What are Analytics Headers?

Requesty supports two kinds of analytics headers:

Origin headers — HTTP-Referer and X-Title identify which app or tool is making the request. Set them once and every request is automatically tagged.
Custom analytics headers — X-Requesty-* headers let you attach arbitrary metadata (agent name, branch, team, environment, etc.) to requests.

Both types are captured by Requesty, stripped before forwarding to the AI provider, and made available as dimensions in your analytics dashboards. They are the simplest way to add metadata — no SDK changes, no request body modifications. Just set an HTTP header.

curl https://router.requesty.ai/v1/chat/completions \
  -H "Authorization: Bearer $REQUESTY_API_KEY" \
  -H "HTTP-Referer: https://yourapp.com" \
  -H "X-Title: My App" \
  -H "X-Requesty-Agent: my-support-bot" \
  -H "X-Requesty-Environment: production" \
  -H "X-Requesty-Team: platform" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-sonnet-4-5",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Origin Headers — `HTTP-Referer` & `X-Title`

Requesty recognizes two standard origin headers that identify where a request comes from:

Header	Stored as	Purpose
`HTTP-Referer`	`origin_referer`	Your site or app URL (e.g., `https://yourapp.com`)
`X-Title`	`origin_title`	A human-readable app name (e.g., `My App`)

These headers are optional but recommended. Add them once via default_headers and every request is tagged automatically. Many integrations set them for you — for example, Cline, Roo Code, Claude Code, and Open WebUI all send their own HTTP-Referer and X-Title values so you can filter traffic by tool in your dashboards.

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_REQUESTY_API_KEY",
    base_url="https://router.requesty.ai/v1",
    default_headers={
        "HTTP-Referer": "https://yourapp.com",
        "X-Title": "My App",
    },
)

In analytics, filter or Group By origin_title to break down usage by app, or by origin_referer to see traffic by URL.

Custom Analytics Headers — `X-Requesty-*`

How It Works

Add one or more X-Requesty-* headers to your request.
Requesty extracts the headers and removes them before forwarding to the AI provider — no data leaks upstream.
The header values are stored as custom metadata fields on the request.
In analytics, filter or Group By those fields to slice your data any way you want.

Header matching is case-insensitive — x-requesty-team, X-REQUESTY-TEAM, and X-Requesty-Team are all treated the same.

Naming Your Headers

Any header matching the X-Requesty-<Name> pattern is captured. The <Name> part (after the prefix) becomes the field key in analytics.

Header	Analytics field
`X-Requesty-Agent`	`Agent`
`X-Requesty-Branch`	`Branch`
`X-Requesty-Environment`	`Environment`
`X-Requesty-Team`	`Team`
`X-Requesty-Customer`	`Customer`

You can use any name you like — there is no fixed list. Pick names that match the dimensions you care about.

Use Cases

Name your AI agent

Give each agent or bot a distinct identity so you can track cost and usage per agent:

-H "X-Requesty-Agent: support-bot"
-H "X-Requesty-Agent: code-reviewer"
-H "X-Requesty-Agent: data-pipeline"

Then in the Advanced analytics tab, Group By → Agent to see a cost and request breakdown per agent.

Track by environment

Separate production traffic from development and staging:

-H "X-Requesty-Environment: production"

Track by team or department

Attribute AI costs to the team that generated them:

-H "X-Requesty-Team: backend"
-H "X-Requesty-Department: engineering"

Track by customer

If you’re building AI features for multiple customers, tag each request:

-H "X-Requesty-Customer: acme-corp"

Track git context

Tag requests with the branch and repo to see which feature branches are driving cost:

-H "X-Requesty-Branch: feat/new-onboarding"
-H "X-Requesty-Repo: myorg/backend"

The Claude Code analytics wrapper does this automatically — it tags every request with your current git branch, repo, agent version, and OS username.

Implementation Examples

Python (OpenAI SDK)

import openai

client = openai.OpenAI(
    api_key="YOUR_REQUESTY_API_KEY",
    base_url="https://router.requesty.ai/v1",
    default_headers={
        "X-Requesty-Agent": "my-support-bot",
        "X-Requesty-Environment": "production",
    }
)

response = client.chat.completions.create(
    model="anthropic/claude-sonnet-4-5",
    messages=[{"role": "user", "content": "Hello"}]
)

Node.js (OpenAI SDK)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_REQUESTY_API_KEY",
  baseURL: "https://router.requesty.ai/v1",
  defaultHeaders: {
    "X-Requesty-Agent": "my-support-bot",
    "X-Requesty-Environment": "production",
  },
});

const response = await client.chat.completions.create({
  model: "anthropic/claude-sonnet-4-5",
  messages: [{ role: "user", content: "Hello" }],
});

cURL

curl https://router.requesty.ai/v1/chat/completions \
  -H "Authorization: Bearer $REQUESTY_API_KEY" \
  -H "X-Requesty-Agent: my-support-bot" \
  -H "X-Requesty-Environment: production" \
  -H "X-Requesty-Team: backend" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4.1",
    "messages": [{"role": "user", "content": "Summarize this document"}]
  }'

Claude Code (automatic)

Install the Requesty analytics wrapper and every Claude Code session automatically sends:

Header	Value
`X-Requesty-Branch`	Current git branch
`X-Requesty-Repo`	`org/repo` from git origin
`X-Requesty-Ai-Agent`	Claude Code version
`X-Requesty-User`	OS username

curl -fsSL https://www.requesty.ai/claude/install.sh | bash

See Claude Code Analytics for details.

Viewing in Dashboards

Once you start sending analytics headers, the fields appear in the Analytics dashboard:

Open the Advanced tab.
In the Group By dropdown, select your custom field (e.g., Agent, Environment, Team).
Choose a Metric (Cost, Requests, Tokens, Latency, etc.) and a Time Range.
Optionally add Filters to narrow down (e.g., Agent = support-bot).

You can combine analytics headers with any other grouping or filter — for example, group by Agent and filter by model = anthropic/* to see which agents use Anthropic models the most.

Establish naming conventions for your analytics headers across your organization. Consistent names like X-Requesty-Team, X-Requesty-Environment, and X-Requesty-Agent make dashboards easy to read for everyone.