API Rate Limits by Platform for Offline Conversion Uploads
Reference table of API rate limits and batch sizes for offline conversion uploads across Google Ads, Meta Conversions API, Microsoft Ads, and TikTok Events API.
When automating offline conversion uploads via API, you need to stay within each platform’s rate limits. Exceeding them results in throttling or errors.
Rate Limit Comparison
| Platform | Events per request | What else is published |
|---|---|---|
| Google Data Manager | 2,000 events | 100,000 ingest writes a day and 300 a minute, per Cloud project |
| Meta Conversions API | 1,000 | One invalid event rejects the whole request |
| Microsoft Ads | 1,000 | Campaign Management API v13 |
| TikTok Events API 2.0 | 1,000 | Send the event as it happens |
Google offline uploads
Send the job through the Data Manager API. One events request holds 2,000 events. A script that instead puts 2,000 booked jobs on the old Google Ads upload call is rejected when the developer token was not already uploading. The 2,000-job batch is the right size for Data Manager and the wrong call for a new token.
The Cloud project is also limited to 100,000 ingest writes a day and 300 a minute. Past either of those, the call comes back as too many requests. The old Google Ads API figure of 15,000 operations a day is the Basic-access cap on that API, not this project limit, and the current quotas page does not list 10 calls per 100 seconds.
Wait 6 hours after the click before you upload it: a click at 9:00 a.m. is too recent at noon and is ready at 3:00 p.m.
Meta Conversions API limits
Pin Graph API v26.0. That is the newest version with a published release date, July 29, 2026. Send at most 1,000 events in one request. One event older than 7 days rejects the whole request, so a batch of 1,000 with one stale booking drops every row.
Microsoft Ads API Limits
| Limit | Value |
|---|---|
| Conversions per request | 1,000 |
| Daily upload limit | Standard API quota |
| Minimum click-to-upload delay | 4 hours |
TikTok Events API Limits
| Limit | Value |
|---|---|
| Events per request | 1,000 |
| Requests above this limit | Rejected |
| Recommended send pattern | As the event occurs |
| Maximum events per request | 1,000 |
How TUOCA Manages Rate Limits
- Automatic batching — Groups events into optimal batch sizes per platform
- Rate throttling — Respects per-platform limits with backoff
- Retry with delay — Rate limit responses trigger automatic retry after delay
- Priority queuing — Meta events processed first due to 7-day window