API Connection Guide

API Limits & Connection Guide

Detailed breakdown of API limits, connection methods, and available data for each service

Last verified: 2026-03-21(verified by: Kenta)

API limits may change without notice

API Limits by Service

A summary of API limits for usage data retrieval used by Konsumpsy.

This table is maintained manually based on official documentation. API limits may change without notice. Always check each service's official documentation for the latest information.

Service Usage Retrieval Required Key Rate Limit Recommended Frequency Notes
Claude API Admin API Key 120 req/min Once daily Known 429 issues with oauth/usage endpoint
Claude Pro Browser usage not accessible
OpenAI API API Key Tier-dependent (5–4,000 RPM) Once daily Tier 1+ recommended
ChatGPT Plus Browser usage not accessible
Vercel Access Token Undisclosed (~60 req/min measured) Once daily
GitHub Copilot Personal Access Token 5,000 req/hour Once daily Org Admin required
Supabase Personal Access Token 120 req/min (standard) Once daily Heavy endpoints: 10 req/min
Cursor No public API
Windsurf No public API
Midjourney No public API
ElevenLabs API Key Undisclosed (~30 req/min measured) Once daily Credit balance only
Perplexity API Key Tier-dependent Once daily
Manus No public API

Fetch Frequency in Konsumpsy

All services are synced in a once-daily batch (UTC 02:00). Real-time retrieval is not performed due to provider rate limits and operational cost considerations.

API Connection Guide by Service

Claude / Anthropic — API Connection Guide

API AvailableVerified: 2026-03-21
Key: Admin API Key
Limit: 120 req/min
Recommended: Once daily

How to Get Your Key

  1. Log in to console.anthropic.com
  2. Left menu → "Settings" → "API Keys"
  3. Click "Create key" → Role: Administrator
  4. Copy the generated key and paste it into Konsumpsy

Important: You need an Admin API Key, not a regular API key. Setting the Role to "Administrator" is required to access the Usage API.

Data Available

Data Available
Monthly token usage
Usage by model (Opus/Sonnet/Haiku)
Daily usage history
Claude.ai (browser) usage
Billing amount ✗ (estimate from tokens only)

Rate Limit Details

The Management API standard rate limit is 120 requests/minute.

Note: The oauth/usage endpoint has a known issue with frequent 429 errors under high access. Konsumpsy therefore performs one batch retrieval per day only.

Permissions

While Admin API Keys have broader permissions than regular keys, Konsumpsy only performs the following operations:

  • ✓ Reading usage data
  • ✗ Sending requests to models
  • ✗ Changing settings
  • ✗ Any other operations

You can revoke the key instantly from Anthropic Console if needed.

View official documentation →

OpenAI — API Connection Guide

API AvailableVerified: 2026-03-21
Key: API Key (sk-proj-...)
Limit: Tier-dependent (5–4,000 RPM)
Recommended: Once daily

How to Get Your Key

  1. Log in to platform.openai.com
  2. Click your avatar (top right) → "API keys"
  3. Click "Create new secret key"
  4. Copy the generated key and paste it into Konsumpsy

Recommended: Check your Usage Tier.

  • Free / Tier 0: RPM 5 (very low) → may affect usage retrieval
  • Tier 1+ ($5+ spend): RPM 500+ → stable retrieval

Data Available

Data Available
Monthly token usage (input / output)
Usage by model
Daily usage history
ChatGPT Plus (browser) conversation usage
Billing amount ✗ (estimate only)

Rate Limit Details

OpenAI rate limits vary by Usage Tier:

Tier RPM TPM
Free 3 40,000
Tier 1 500 200,000
Tier 2 5,000 2,000,000
Tier 3+ 5,000+ 4,000,000+

Konsumpsy accesses the API once per day, so Tier 1+ is sufficient.

View official documentation →

Vercel — API Connection Guide

API AvailableVerified: 2026-03-21
Key: Vercel Access Token
Limit: Undisclosed (~60 req/min measured)
Recommended: Once daily

How to Get Your Token

  1. Go to vercel.com/account/tokens
  2. Click "Create Token" → Scope: Full Account
  3. Copy the generated token and paste it into Konsumpsy

Data Available

Data Available
Number of projects and deployments
Bandwidth usage
Serverless Function execution time

Rate Limit Details

Vercel does not publicly disclose its API rate limits. Measured in practice at approximately 60 requests/minute.

Konsumpsy accesses the API once per day to stay within safe limits.

View official documentation →

Supabase — API Connection Guide

API AvailableVerified: 2026-03-21
Key: Personal Access Token
Limit: 120 req/min (standard) / 10 req/min (heavy)
Recommended: Once daily

How to Get Your Token

  1. Go to app.supabase.com/account/tokens
  2. Click "Generate new token" and enter a name (e.g., Konsumpsy)
  3. Copy the generated token and paste it into Konsumpsy

Data Available

Data Available
Number of projects and status
Database usage (connections, storage)
API request count (via Management API)

Rate Limit Details

Endpoint Rate Limit
Standard endpoints 120 req/min
Heavy endpoints (analytics, etc.) 10 req/min

Konsumpsy performs one batch retrieval per day and will not reach these limits.

View official documentation →

GitHub / Copilot — API Connection Guide

API AvailableVerified: 2026-03-21
Key: Personal Access Token (Classic)
Limit: 5,000 req/hour
Recommended: Once daily

How to Get Your Token

  1. Go to github.com/settings/tokens
  2. Click "Generate new token (classic)"
  3. Scope: check read:org (required for Copilot usage retrieval)
  4. Copy the generated token and paste it into Konsumpsy

Important: Organization Admin permissions are required to retrieve Copilot usage. Personal Copilot plans do not have a usage API — only seat count is available.

Data Available

Data Available
Copilot seat count (org)
Active user count
Code completion activity (aggregated)

Rate Limit Details

With authentication: 5,000 requests/hour (very generous)

Konsumpsy calls the /orgs/{org}/copilot/usage endpoint once per day only.

View official documentation →

ElevenLabs — API Connection Guide

API AvailableVerified: 2026-03-21
Key: API Key
Limit: Undisclosed (~30 req/min measured)
Recommended: Once daily

How to Get Your Key

  1. Go to elevenlabs.io/app/settings/api-keys
  2. Click "Create API Key"
  3. Copy the generated key and paste it into Konsumpsy

Data Available

Data Available
Credit balance (used / limit)
Subscription plan name
Next renewal date
Credit usage breakdown

Rate Limit Details

ElevenLabs does not publicly disclose its API rate limits. Measured in practice at approximately 30 requests/minute.

Konsumpsy calls the /v1/user/subscription endpoint once per day only.

View official documentation →

Perplexity — API Connection Guide

API AvailableVerified: 2026-03-21
Key: API Key
Limit: Tier-dependent
Recommended: Once daily

How to Get Your Key

  1. Go to perplexity.ai/settings/api
  2. Create an API key
  3. Copy the generated key and paste it into Konsumpsy

Data Available

Data Available
API usage
Credit balance
Browser usage

Rate Limit Details

Perplexity API rate limits vary by account Tier.

Konsumpsy performs one batch retrieval per day and will not reach rate limits under normal use.

View official documentation →

Midjourney

No APIVerified: 2026-03-21

No Public API

Midjourney does not currently provide a public API for usage data retrieval.

What Konsumpsy Can Do

  • ✓ Manual entry of plan, cost, and renewal date
  • ✓ Invoice upload for Actual History (Pro plan only)
  • ✓ AI recommendations (based on renewal date when usage data is unavailable)
  • ✗ Automatic usage retrieval

Konsumpsy will add support promptly if Midjourney releases a public API.

Cursor

No APIVerified: 2026-03-21

No Public API

Cursor does not currently provide a public API for usage data retrieval.

What Konsumpsy Can Do

  • ✓ Manual entry of plan, cost, and renewal date
  • ✓ Invoice upload for Actual History (Pro plan only)
  • ✓ AI recommendations (based on renewal date when usage data is unavailable)
  • ✗ Automatic usage retrieval

Konsumpsy will add support promptly if Cursor releases a public API.

Windsurf

No APIVerified: 2026-03-21

No Public API

Windsurf does not currently provide a public API for usage data retrieval.

What Konsumpsy Can Do

  • ✓ Manual entry of plan, cost, and renewal date
  • ✓ Invoice upload for Actual History (Pro plan only)
  • ✓ AI recommendations (based on renewal date when usage data is unavailable)
  • ✗ Automatic usage retrieval

Konsumpsy will add support promptly if Windsurf releases a public API.

Notes on Interpreting Limits

1

"Can't retrieve usage" vs "Usage is 0" are different

For services without an API (e.g., Midjourney), the usage field shows "No data". This does not mean the service is unused.

2

Data freshness

Usage data for all services is updated once per day. For real-time usage, check each service's own dashboard directly.

3

Token count vs billed amount

The "cost estimate" calculated from API token counts may differ from your actual bill due to promotional credits, volume discounts, and currency conversion rates. For accurate billing, supplement with the invoice upload feature on the Pro plan.

4

API key permissions

Konsumpsy uses API keys for read-only usage data access. We recommend granting minimum necessary permissions to each service.

5

Rate limit errors (429)

Konsumpsy retries automatically with exponential backoff. If retrieval still fails, the previous value is shown with a "Data refresh failed" badge. Automatic recovery occurs on the next daily run.

FAQ

Q. What can Konsumpsy do after I register my API key?

A. Read-only access to usage data. We never send requests to the service, change settings, or initiate cancellations.

Q. How is my API key stored?

A. Keys are encrypted with AES-256 before being stored in the database. They are never stored or displayed in plaintext.

Q. Can I change how often usage data is fetched?

A. Currently fixed at once per day. We are exploring options to make this configurable in a future update.

Q. What should I do if my API key is compromised?

A. Immediately revoke the key from the service's dashboard, then remove and re-register it in Konsumpsy's service settings.

Q. Can I add a service that isn't listed?

A. Yes — you can manually register service name, monthly cost, and renewal date. Automatic usage retrieval won't work, but cost management and renewal alerts will.