Connect Requesty

Requesty is an LLM gateway that routes OpenAI- and Anthropic-compatible requests to hundreds of models across many providers. The Requesty integration imports your organization's daily spend and token usage into Cloud Intelligence™ so you can analyze LLM spend in Cloud Analytics by model, hosting provider, API key, and routing policy.

Prerequisites

• A Requesty organization in which you can create API keys with Manage permission.

• A DoiT account with the Billing Profiles Admin permission.

Required credentials

Field	Description	Where to find it
Organization name	A short name for the Requesty organization, used to label the connection. Lowercase letters, digits, and hyphens only	Your choice
Management API key	An API key with Manage permission set to Read	Requesty Console > API Keys

Limit the key's spend

Requesty's usage API only accepts keys with Manage permission. Keys without it are rejected. As a safeguard, also set a small monthly spend limit on the key you share with DoiT. DoiT only reads usage and never sends inference requests.

Create a management API key

1. Sign in to the Requesty Console and navigate to API Keys.

2. Select Create New API Key and give it a descriptive name, for example doit-cloud-intelligence.

3. Set Manage permission to Read and Completions permission to None.

4. Set a monthly spend limit of 0.01.

5. Copy the key.

   注意

   The key is only shown once. Make sure to copy it before closing the dialog.

Connect your Requesty organization

1. Sign in to Cloud Intelligence, select Data ingestion and integrations > Integrations from the top navigation mega menu.

2. Search for Requesty in the integrations catalog.

3. Select Connect + on the integration card.

4. Enter the Organization name and paste the Management API key.

5. Select Test connection to verify DoiT can read your Requesty usage.

6. If the test is successful, select Save to complete the setup.

   If the test fails or you need further assistance, please open an expert inquiry.

注意

Upon initial connection, DoiT imports up to 12 months of historical Requesty usage. The import process can take a few hours. You'll receive an email when the data is available in Cloud Analytics.

How spend becomes cost

Requesty draws each request's model cost, at the provider's list price, from your prepaid credit balance, and charges a 5% markup when you buy credits.

• Cost is the credits drawn plus the 5% markup, so it matches what you pay for credits.

• List cost is the credits drawn at provider list price, as shown in the Requesty analytics dashboard.

• Bring your own keys (BYOK) requests are billed by the upstream provider directly, so they carry zero cost in Requesty data to avoid double counting your provider bill. Their estimated provider spend is available as the genai.cost.byok_external metric.

注意

Failed requests appear in request counts with zero cost. Prompt-cache tokens are included in input tokens, because Requesty does not report them separately.

Data refresh

DoiT refreshes your Requesty usage every 6 hours. Requesty reports usage at daily granularity in UTC, and a day is imported once it has closed.

Reporting dimensions

Once connected, Requesty data maps to the following standard Cloud Analytics dimensions:

Cloud Analytics dimension	Requesty value	Example
Provider	Requesty	Requesty
Billing Account	Organization name	my-org
Project	API key ID	9f2c…
Service	Model inference	Model Inference
SKU	Hosting provider and model	openai/gpt-4.1-nano, bedrock/claude-haiku-4-5@us-east-1

In addition to standard dimensions, the following Requesty-specific labels are available:

Label	Description
Model	The hosting provider and model that served the request
Provider	The hosting provider that served the request, for example openai, azure, bedrock, or vertex
Routing policy	The routing policy the request was sent to, when the model was chosen by a fallback, load-balancing, or latency policy
API key	The ID of the API key that made the requests
BYOK	true when the request used your own provider key
Organization	The Requesty organization the charge belongs to

Usage amounts are measured in tokens. Input tokens, output tokens, and request counts are available as the token.usage.input, token.usage.output, and genai.usage.requests metrics.

注意

Requesty costs appear in the GenAI Intelligence dashboard alongside other supported GenAI providers.

Delete your Requesty connection

1. Sign in to Cloud Intelligence, select Data ingestion and integrations > Integrations from the top navigation mega menu.

2. Search for Requesty in the integrations catalog.

3. Select Manage on the integration card.

4. Select the kebab menu (⋮) at the rightmost end of the organization you want to remove and select Delete.

5. Select Delete to confirm.