Bulk Create Geofences from Addresses

The bulk-create API is the integration entry point for syncing address-based geofences into SpatialFlow from your existing systems. Each address is geocoded and turned into a circular geofence in a single call, no GeoJSON required.

When to Use This

Use this endpoint when you need to:

Sync from a CRM or TMS: import customer sites, service zones, or delivery stops from Salesforce, HubSpot, Onfleet, or any flat address list
Onboard a new customer: turn a customer's address list into geofences in one API call
Run a scheduled refresh: re-import a CSV nightly; dedup_strategy: "skip" makes re-imports idempotent (duplicates return a match report, not an error)
One-shot CSV import: a team member pastes addresses into a script and fires a single POST to get all fences created

For interactive imports with a preview step, use the /api/v1/geofences/preview endpoint before committing.

Authentication

API keys are recommended for integrations. API key calls get a 60 requests/hour limit per workspace (6× the JWT limit), and they don't expire between user sessions.

Generate an API key from Settings > API Keys, then include it in the X-API-KEY header:

curl https://api.spatialflow.io/api/v1/geofences/bulk \
  -H "X-API-KEY: sf_..." \
  -H "Content-Type: application/json" \
  -d '{"items": [{"address": "123 Main St, San Francisco, CA"}]}'

JWT tokens also work. Pass them in the Authorization: Bearer <token> header. An API key needs the geofences:create permission.

Request Shape

POST /api/v1/geofences/bulk

{
  "items": [
    {
      "address": "123 Main St, San Francisco, CA",
      "name": "SF Downtown Hub",
      "buffer_meters": 100,
      "tags": ["customer:acme", "tier:gold"]
    },
    {
      "address": "456 Market St, San Francisco, CA"
    }
  ],
  "dedup_strategy": "skip"
}

Item fields

Field	Type	Required	Description
`address`	string	Yes	Street address to geocode (max 500 chars)
`name`	string	No	Display name (max 255 chars). Derived from address if omitted.
`buffer_meters`	int	No	Fence radius in meters (1 to 10000, default 100)
`tags`	string[]	No	Up to 20 tag strings per item

dedup_strategy

Controls what happens when a geocoded point falls within an existing geofence's buffer zone:

Strategy	Behavior
`skip` (default)	Duplicates return `status: "duplicate"` with a `dedup_match` object. Nothing is written. Re-submitting the same batch is idempotent.
`override`	Duplicates are created anyway. Results carry `dedup_match` so you know which were forced. Use this for intentional dual-fencing of the same location.
`fail`	First duplicate aborts the entire batch. HTTP 409 is returned with the partial result list. Use this for strict transactional flows.

Response Shape

All three strategies return the same response envelope. HTTP 201 means the batch was processed; some items may still have status: "error" or status: "duplicate". Check each result individually.

{
  "results": [
    {
      "index": 0,
      "status": "created",
      "geofence_id": "550e8400-e29b-41d4-a716-446655440000"
    },
    {
      "index": 1,
      "status": "duplicate",
      "dedup_match": {
        "geofence_id": "660e8400-e29b-41d4-a716-446655440001",
        "name": "Existing SF Hub",
        "distance_meters": 12.3
      }
    },
    {
      "index": 2,
      "status": "error",
      "error_code": "GEOCODE_FAILED",
      "error": {
        "message": "Could not geocode address",
        "code": "GEOCODE_FAILED"
      }
    }
  ]
}

index is 0-based and matches the position of the item in the request items array. For error results, error_code is the stable machine-readable key and mirrors the nested error.code field retained for existing clients.

HTTP status codes

HTTP	Meaning
201	Batch processed (mixed results possible; check each `results[i].status`)
401	Missing or invalid auth
403	Workspace lacks the BULK_IMPORT feature flag
409	`dedup_strategy: "fail"` triggered; partial results in `details.results`
413	Batch exceeds workspace cap (see Batch Size Cap)
422	Request body invalid (malformed JSON, missing required fields)
429	Rate limit exceeded (see Rate Limits)

Per-item error codes

Code	Cause
`GEOCODE_FAILED`	Geocoder could not match the address. Check spelling and include city/state/ZIP.
`GEOCODER_RATE_LIMIT`	The geocoder hit an upstream rate limit mid-batch. Retry the failed items after a short delay.
`INVALID_TAGS`	Tags failed validation: too long, invalid characters, or more than 20 per item.
`CREATE_FAILED`	Catch-all server error during geofence creation (rare). Retry the specific item.

curl Example

Minimal happy path with two addresses. The first is created, the second is a near-duplicate that triggers the default skip strategy:

curl -s -X POST https://api.spatialflow.io/api/v1/geofences/bulk \
  -H "X-API-KEY: sf_..." \
  -H "Content-Type: application/json" \
  -d '{
    "items": [
      {
        "address": "1 Infinite Loop, Cupertino, CA",
        "name": "Apple HQ",
        "buffer_meters": 200,
        "tags": ["customer:apple", "tier:enterprise"]
      },
      {
        "address": "1 Apple Park Way, Cupertino, CA",
        "name": "Apple Park"
      }
    ]
  }'

Example response:

{
  "results": [
    {
      "index": 0,
      "status": "created",
      "geofence_id": "550e8400-e29b-41d4-a716-446655440000"
    },
    {
      "index": 1,
      "status": "duplicate",
      "dedup_match": {
        "geofence_id": "550e8400-e29b-41d4-a716-446655440000",
        "name": "Apple HQ",
        "distance_meters": 483.2
      }
    }
  ]
}

Python SDK Example

The most common integration pattern is a two-pass import: detect duplicates with dedup_strategy: "skip", then force-create only the ones you intentionally want to dual-fence with dedup_strategy: "override".

import asyncio
from spatialflow import SpatialFlow
from spatialflow._generated.spatialflow_generated.models import (
    BulkItemInput,
    BulkCreateRequest,
)


async def sync_customer_sites(addresses: list[dict]) -> None:
    async with SpatialFlow(api_key="sf_...") as client:
        items = [
            BulkItemInput(
                address=row["address"],
                name=row.get("name"),
                tags=row.get("tags", []),
            )
            for row in addresses
        ]

        # First pass: detect duplicates without writing
        first_pass = await client.geofences.bulk_create(
            BulkCreateRequest(items=items, dedup_strategy="skip")
        )

        created = [r for r in first_pass.results if r.status == "created"]
        duplicates = [r for r in first_pass.results if r.status == "duplicate"]
        errors = [r for r in first_pass.results if r.status == "error"]

        print(f"First pass: {len(created)} created, {len(duplicates)} duplicate, {len(errors)} error")

        # Log errors: these need address corrections before re-import
        for r in errors:
            print(f"  Error at index {r.index}: {r.error.message} ({r.error.code})")

        # Second pass: force-create duplicates the caller explicitly wants to override
        # (e.g. a customer opened a second location at the same address)
        duplicates_to_force = [items[r.index] for r in duplicates]
        if duplicates_to_force:
            override_pass = await client.geofences.bulk_create(
                BulkCreateRequest(
                    items=duplicates_to_force,
                    dedup_strategy="override",
                )
            )
            forced = [r for r in override_pass.results if r.status == "created"]
            print(f"Override pass: {len(forced)} force-created")

        # Collect all created geofence IDs
        all_created_ids = [r.geofence_id for r in first_pass.results if r.status == "created"]
        print(f"Total geofences created: {len(all_created_ids)}")


if __name__ == "__main__":
    customer_sites = [
        {"address": "123 Main St, San Francisco, CA", "tags": ["customer:acme"]},
        {"address": "456 Market St, San Francisco, CA", "tags": ["customer:acme"]},
        {"address": "789 Mission St, San Francisco, CA", "tags": ["customer:acme"]},
    ]
    asyncio.run(sync_customer_sites(customer_sites))

Install the SDK:

pip install spatialflow

The BulkItemInput and BulkCreateRequest classes are generated from the OpenAPI spec and available in spatialflow._generated.spatialflow_generated.models. The client.geofences.bulk_create() method is the typed wrapper in GeofencesResource.

Rate Limits

The /bulk endpoint has dedicated rate limits separate from the general API limits:

Auth Method	Limit
JWT (browser session)	10 requests/hour per workspace
API key	60 requests/hour per workspace

With the default batch size cap of 500 items, an API key integration can create up to 30,000 geofences/hour (500 items × 60 requests/hour).

When the rate limit is exceeded, the API returns HTTP 429 with a Retry-After header indicating when you can retry:

{
  "detail": "Rate limit exceeded. Try again in 3412 seconds.",
  "error_code": "RATE_LIMIT_EXCEEDED"
}

Batch Size Cap

Each workspace has a batch size limit (default 500 items, ceiling 1000), stored per workspace but with no dashboard control to change it yet. Email support@spatialflow.io to raise it within the 500-to-1000 range; going above 1000 needs SpatialFlow ops to intervene directly.

Requests above the cap are rejected with HTTP 413, and no items are processed:

{
  "detail": "Batch too large: 501 > 500",
  "error_code": "BATCH_TOO_LARGE",
  "details": {
    "max_items": 500,
    "requested_items": 501
  }
}

Split large imports into chunks on the client side and send them sequentially or with a short delay between batches to respect rate limits.

Other SDKs

Node SDK: client.geofences.appsGeofencesApiBulkCreateGeofences({ bulkCreateRequest: { items, dedup_strategy } }). See the spatialflow-node repository.
Go SDK: client.BulkCreateGeofences(ctx, spatialflow.BulkCreateRequest{ Items: items, DedupStrategy: "skip" }). See the spatialflow-go repository.

Both SDKs handle authentication and error mapping. Neither retries a bulk create: the Go client retries only idempotent methods, and the Node client doesn't retry. Retry a 429 yourself after its Retry-After delay, and a 5xx with backoff. With dedup_strategy: "skip", resubmitting the same batch is idempotent.

OpenAPI Spec

The complete schema, including the BulkCreateRequest, BulkCreateResponse, BulkResult and BulkItemInput definitions, is in the OpenAPI document that this reference is built from. To generate a client, save a copy of that file with your code, so a later API release doesn't change your client unexpectedly:

curl -o spatialflow-openapi.json https://docs.spatialflow.io/openapi/openapi.json
openapi-generator generate \
  -i spatialflow-openapi.json \
  -g python \
  -o ./my-spatialflow-client

When to Use This​

Authentication​

Request Shape​

Item fields​

dedup_strategy​

Response Shape​

HTTP status codes​

Per-item error codes​

curl Example​

Python SDK Example​

Rate Limits​

Batch Size Cap​

Other SDKs​

OpenAPI Spec​