🔑

API Key Required

Authorization Gate

Enter your tenant API key to open tenant index operations.

This page covers all 12 index management operations. Each operation shows a live curl request preview and the raw server JSON response so you can integrate directly. Copy any of the curl requests below directly into your terminal to reproduce a call.
Tenant ID
Indexed Documents 0
Geo Index No
Custom Fields None
Access Scope
Operations are listed in recommended order: Read schema → Browse documents → Upsert → Delete → Increment → Bulk upload → Synonyms → Compact → Bulk Import (Async) → Export Snapshot → Import Snapshot → Drop Index. Each operation is independent — skip directly to what you need.
1

GET /api/tenant/schema

Read Tenant Index Schema

Validates the API key and returns tenant metadata: document count, geo availability, and a sample of custom field names present in the index.

Run this first after logging in to confirm your key works and to check how many documents are currently indexed.

Operation Result

Run this operation to see the server JSON response.
2

GET /api/index/docs

Browse Indexed Documents

Lists documents in the tenant index as JSON, paginated server-side. Max 100 documents per page.

Use this to inspect what is actually stored in the index. Useful for debugging before running searches.
IDNameGeoFields JSON
Load documents to browse the index.

Page 1

JSON Response

Run this operation to see the server JSON response.
3

POST /api/index/doc

Upsert a Document

Creates or replaces a document by ID. To create the geo index, include both lat and lon in a document — there is no separate geo-index endpoint.

Upsert = insert or update. If a document with the same ID already exists, it is fully replaced. Custom metadata fields support string, number, and boolean types — numbers enable numeric filtering and faceting in search.

Unique identifier — used for updates and deletion

The primary text field indexed for search

Geo Document Mode

Enable when the document represents a location (place, store, delivery point). Both coordinates are required; the first geo document activates the tenant's spatial index.

Geo mode is off. The document will be indexed without location unless both coordinate fields are filled.

Add structured data fields: numbers enable numeric filters, booleans enable boolean filters. Empty rows are ignored.

Vector Embedding (optional)

Powers Hybrid Search — a search mode that combines this document's text with a semantic vector, using an HNSW nearest-neighbor index. Paste a list of numbers produced by your own embedding model (e.g. from OpenAI, Cohere, or a local model). Every document across this tenant must use vectors of the same length. Leave empty if you don't need semantic search.

No vector set — this document will not be found by Hybrid/Semantic Search, only by regular keyword search.

Payload Preview


                

Operation Result

Run this operation to see the server JSON response.
4

DELETE /api/index/doc/{id}

Delete a Document

Removes one document by its ID and writes the delete operation to the activity log.

Deletion is permanent and cannot be undone from this UI. The document ID must match exactly.

Operation Result

Run this operation to see the server JSON response.
5

POST /api/index/doc/{id}/increment

Atomic Field Increment

Atomically increments a numeric custom field for an existing document without a full re-index.

Use case: Track popularity, view_count, or click_count fields without re-uploading the whole document. Safe under concurrent access.

Must be an existing numeric custom field

Negative values decrement the field

Operation Result

Run this operation to see the server JSON response.
6

POST /api/index/upload

Bulk Import Documents

Upload a .json or .jsonl file. A JSON file may contain a single document or an array.

JSONL format: One JSON object per line — ideal for large datasets. JSON array: [{…},{…}] — works for small batches. Large uploads may take a few seconds while the backend indexes records.

Operation Result

Run this operation to see the server JSON response.
7

GET / POST / DELETE /api/index/synonyms

Manage Tenant Synonyms

Synonym groups expand lookup terms for this tenant only. Searching for any term in a group returns results indexed with any of the group's terms.

Example: Set source term = mobile and synonyms = phone, iphone, smartphone. A user searching for "phone" will now also find documents indexed with the word "mobile".
No tenant synonyms loaded. Click List Synonyms to fetch.
List synonyms
Add / update synonyms
Delete synonym term

Operation Result

Run this operation to see the server JSON response.
8

POST /api/index/compact

Compact Activity Log

Saves the current index state to a snapshot file and compacts the activity log to reduce disk usage.

Compaction is safe and non-destructive but rewrites the snapshot file. Run it periodically after large batch imports to keep the activity log small. Avoid running during active high-write periods.

Operation Result

Run this operation to see the server JSON response.
9

POST /api/index/bulk_import

Bulk Import (Async, Large Datasets)

Submits a JSON array of documents as a background job and returns a job ID immediately. Indexing continues after the response — poll for progress below. Job state lives in shared storage, so any node in a cluster can report progress.

Unlike Bulk Import Documents above (Operation 6), which blocks until the whole file is indexed, this operation returns instantly and indexes in the background — better for very large datasets. Paste a JSON array of documents (each needs at least id; name, latitude/longitude, and any custom fields are optional).

No job running.

Start job
Poll progress

Operation Result

Run this operation to see the server JSON response.
10

GET /api/index/export

Export Snapshot (Backup)

Compacts the index and downloads the full tenant snapshot as a binary file — documents, synonyms, ranking config, and schema.

Use this before risky changes, or on a schedule, to keep a restorable backup of this tenant's index. The downloaded file can be restored with the Import Snapshot operation below.

Operation Result

Run this operation to see the server JSON response.
11

POST /api/index/import

Import Snapshot (Restore)

Uploads a binary snapshot file and fully replaces this tenant's current index — documents, synonyms, ranking config, and schema.

Destructive: this replaces the entire current tenant index. Only import a snapshot file downloaded from the Export Snapshot operation.

Operation Result

Run this operation to see the server JSON response.
12

POST /api/index/drop

Drop Index

Removes this tenant's index from search entirely. Your API key stays valid — the next document you add starts a brand new, empty index.

Not instantly recoverable by you: once dropped, this index no longer appears anywhere in your account — browsing, searching, and schema all show empty/zero. The underlying files are archived (renamed, not deleted) so an admin can still see that this happened and when, but restoring it requires contacting an admin.

Must exactly match the Tenant ID in the summary bar at the top of this page.

Operation Result

Run this operation to see the server JSON response.

Activity Log

Live
All operations performed on this page appear here with timestamps and server responses.
Ready.