Application, Record, and Index limits
Limit | Value |
---|---|
Application size * | 100 GB (or 1 GB for the Build plan) |
Index size * | 100 GB (or 1 GB for the Build plan) |
Indexing operations | Depends on your plan and usage. |
Indexing rate | 10,000 indexing operations per Unit (as applicable) |
Indexing request body size | 1 GB per batch |
Number of Indices per Application | 1,000 (Premium), 50 (Grow), or 10 (Build) |
Number of custom character normalizations | 10 |
Pending requests | 5,000 |
Record size * | 10 KB to 100 KB—based on your plan (10 KB maximum for the Build plan) |
Search request body size | 50 MB per request |
Synonyms per Index | 10,000 (or 1,000 for the Build plan) |
Throttling limit | 100 pending requests |
Virtual replicas per Index | 20 |
Filters, Facets, and Rules limits
Limit | Value |
---|---|
Conditions per rule | 25 |
Filter score | 65,535 |
Maximum ratio of unique Facet-Value pairs per Record per Index | 0.1 |
Number of Filters | 1,000 |
Promoted items per promote group | 100 |
Promoted items per rule | 300 |
Rule consequence size | 100 KiB |
Size of each query parameters in the consequence | 16 KiB |
Values per Facet per Search query | 1,000 |
Search user interface limits
Limit | Value |
---|---|
Pagination | 20,000 |
Query Suggestions | 100 |
Query size | 512 bytes |
Insights and analytics limits
Limit | Value |
---|---|
Analytics API calls | 100 API calls per minute per app |
Tag size | 100 characters |
Tags processed per Index | 1,750 unique tag combinations every 5 minutes |
Top searches | 1,000 |
Security limits
Limit | Value |
---|---|
Number of API keys | 5,000 |
Connectors limits
Limit | Value |
---|---|
Maximum number of authentications per appID | 200 |
Maximum number of sources per appID | 100 |
Maximum number of destinations per appID | 500 |
Maximum number of tasks bound to a single source per appID | 10 |
Push endpoint maximum payload size | 5 MB per API call |
Push endpoint maximum watch duration | 3 minutes |
Maximum file size for a JSON connector | 1 GB |
Maximum number of rows in file for a JSON connector | 1,000,000 |
Maximum file size for a CSV connector | 1 GB |
Maximum number of rows in file for a CSV connector | 1,000,000 |
Rate limits for connectors API | Value |
---|---|
Push endpoint API | 750 calls per minute per IP |
Observability API (/1/runs) | 200 calls per 30s per IP |
Run task endpoint (/1/tasks/:taskID/run) | 10 calls per 10 minute per taskID |
All other endpoints | 200 calls per minute per IP |
Data transformation and Fetch limits
Limit | Value |
---|---|
Data transformation runtime | 1 second per transformation |
Data transformation memory | 100 kB per request |
Number of transformations per Application | 5 million per month |
Fetch enrichment requests to the Search API | Up to 1 million Fetch requests per Application per month, |
as part the 5 million transformations per Application per month limit. | |
Fetch timeout | 15 seconds per Record |
Fetch response size | 128 kB |
Collections limits
Limit | Value |
---|---|
Number of collections per index | 500 (Premium plan) or 1,000 (Elevate plan) |
Number of conditions per collection | 50 |
Number of manually selected items per collection | 10,000 |
ObjectID size (manually selected items) | 200 characters |
Smart Groups and Composition API
Limit | Value |
---|---|
Number of groups per curated query | 3 |
Number of records per group | 50 |
Highest start position of each group | 100 |
Number of compositions per application | 100 |
Number of composition rules per composition | 200 |
Number of disjunctive facets in a single request | 20 |
DocSearch plan limits
Limit | Value |
---|---|
Application size | 25 GB |
Index size | 25 GB |
Record size | 100 kB |
Number of indices per application | 20 |
Queries per second | 3 |
Maximum number of team members | 10 |
Generative Experience Guides limits
Limit | Value |
---|---|
Maximum number of Guides per Index | 1,000 |
Maximum size per Guide | 100 kB |
Maximum input tokens per generation (based on third-party LLM OpenAI calculation) | 50,000 |
Maximum output tokens per generation (based on third-party LLM OpenAI calculation) | 2,000 |