# Gemini-embedding-001 api issues **URL:** \ **Category:** Gemini API\ **Tags:** gemini-api\ **Created:** [September 14, 2025, 9:44pm UTC](https://discuss.ai.google.dev/t/gemini-embedding-001-api-issues/104597 "2025-09-14T21:44:34Z")\ **Posts on this page:** 3\ **Page:** 1
**Author:** ![Gene](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/gene/32/34397_2.png) [@Gene](https://discuss.ai.google.dev/u/Gene)\ **Post date:** [September 14, 2025, 9:44pm UTC](https://discuss.ai.google.dev/t/gemini-embedding-001-api-issues/104597/1 "2025-09-14T21:44:34Z")
the documentation says “You can also generate embeddings for multiple chunks at once by passing them in as a list of strings.” ```auto curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-embedding-001:embedContent" \ -H "x-goog-api-key: $GEMINI_API_KEY" \ -H 'Content-Type: application/json' \ -d '{"model": "models/gemini-embedding-001", "content": [ {"parts": [{"text": "What is the meaning of life?"}]}, {"parts": [{"text": "What is the purpose of existence?"}]}, {"parts": [{"text": "How do I bake a cake?"}]} ] }' ``` however the api returns this error ```auto { "error": { "code": 400, "message": "Invalid JSON payload received. Unknown name \"content\": Proto field is not repeating, cannot start list.", "status": "INVALID_ARGUMENT", "details": [ { "@type": "type.googleapis.com/google.rpc.BadRequest", "fieldViolations": [ { "description": "Invalid JSON payload received. Unknown name \"content\": Proto field is not repeating, cannot start list." } ] } ] } } ``` ---
**Author:** ![Lalit\_Kumar](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/lalit_kumar/32/37058_2.png) [@Lalit\_Kumar](https://discuss.ai.google.dev/u/Lalit_Kumar)\ **Post date:** [September 15, 2025, 9:51am UTC](https://discuss.ai.google.dev/t/gemini-embedding-001-api-issues/104597/3 "2025-09-15T09:51:07Z")
Hello, Welcome to the Forum! Thank you for highlighting this issue. We were able to reproduce it and it does appear to be a bug. In the meantime, you can refer to the command in the [Gemini API reference](https://ai.google.dev/api/embeddings#example-request_1). ---
**Author:** ![erwinyonata](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/erwinyonata/32/151317_2.png) [@erwinyonata](https://discuss.ai.google.dev/u/erwinyonata)\ **Post date:** [September 20, 2026, 5:40am UTC](https://discuss.ai.google.dev/t/gemini-embedding-001-api-issues/104597/5 "2026-09-20T05:40:17Z")
@Lalit_Kumar is this alrady fixed? I am also had a problem with gemini-embedding-001 batch file using .jsonl file upload through GCS. here my problem ```json { "key": "id_1", "request": { "content": { "parts": [ { "text": "dummy embedding text number 0" } ] }, "embed_content_config": { "output_dimensionality": 768, "task_type": "RETRIEVAL_DOCUMENT", "title": "some_title_1" } }, "status": "Bad Request: {\"error\": {\"code\": 400, \"message\": \"Invalid JSON payload received. Unknown name \\\"content\\\": Cannot find field.\\nInvalid JSON payload received. Unknown name \\\"embed_content_config\\\": Cannot find field.\", \"status\": \"INVALID_ARGUMENT\", \"details\": [{\"@type\": \"type.googleapis.com/google.rpc.BadRequest\", \"fieldViolations\": [{\"description\": \"Invalid JSON payload received. Unknown name \\\"content\\\": Cannot find field.\"}, {\"description\": \"Invalid JSON payload received. Unknown name \\\"embed_content_config\\\": Cannot find field.\"}]}]}}", "response": {}, "processed_time": "2026-09-20T02:11:32.59+00:00" } ``` I already copy and paste the exact payload example for gemini-embedding-001 : [Get batch embeddings inferences  |  Gemini Enterprise Agent Platform  |  Google Cloud Documentation](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/embeddings/batch-prediction-genai-embeddings#gemini-embedding-batch-prediction-preview) and re-run the process, upload to GCS, run batch inference (I can see on google cloud console) then wait, then error, here my code ```python from __future__ import annotations import os import sys import time import asyncio import argparse from datetime import datetime from pathlib import Path sys.path.insert(0, str(Path( __file__ ).resolve().parents[3])) from dotenv import load_dotenv load_dotenv() # populates os.environ for GOOGLE_APPLICATION_CREDENTIALS from google import genai from google.cloud import storage from google.genai import types from settings import settings # basically my settings, you can change the credential to reproduce GEMINI_DIR = Path( __file__ ).resolve().parent POLL_INTERVAL_S = 10 BATCH_INPUT_PATHS = { 'gemini-embedding-001': GEMINI_DIR / 'batching_input_gemini_embedding_001.jsonl', 'gemini-embedding-2': GEMINI_DIR / 'batching_input_gemini_embedding_2.jsonl', } # Input/output GCS paths are derived from the local input file's own # name (not a separate hardcoded URI per model) so the local file, its # GCS upload, and its output directory all share one name to track by. GCS_BUCKET = 'gs://testing_aioverview' _TERMINAL_STATES = { types.JobState.JOB_STATE_SUCCEEDED, types.JobState.JOB_STATE_FAILED, types.JobState.JOB_STATE_CANCELLED, types.JobState.JOB_STATE_PARTIALLY_SUCCEEDED, types.JobState.JOB_STATE_EXPIRED, } def _upload_to_gcs(path: Path, uri: str) -> None: storage_client = storage.Client(project=settings.GOOGLE_CLOUD_PROJECT) blob = storage.Blob.from_uri(uri, client=storage_client) blob.upload_from_filename(str(path)) def _count_rows(path: Path) -> int: with path.open(encoding='utf-8') as f: return sum(1 for line in f if line.strip()) async def _poll_and_report( batches_client, job, *, model: str, input_path: Path, row_count: int, start: datetime, start_monotonic: float, ) -> None: while job.state not in _TERMINAL_STATES: await asyncio.sleep(POLL_INTERVAL_S) job = await batches_client.get(name=job.name) print(f' polled: state={job.state}') end_monotonic = time.monotonic() end = datetime.now() latency_ms = (end_monotonic - start_monotonic) * 1000 print(f'model: {model}') print(f'input file: {input_path} ({row_count} row(s))') print(f'start time: {start.isoformat()}') print(f'end time: {end.isoformat()}') print(f'latency: {latency_ms:.2f} ms') print(f'final state: {job.state}') if job.error: print(f'job error: {job.error}') if job.dest: print(f'output: {job.dest}') async def run_vertex_batch_job(model: str, location: str) -> None: input_path = BATCH_INPUT_PATHS[model] input_gcs_uri = f'{GCS_BUCKET}/{input_path.name}' output_gcs_uri = f'{GCS_BUCKET}/{input_path.stem}_output/' row_count = _count_rows(input_path) await asyncio.to_thread(_upload_to_gcs, input_path, input_gcs_uri) print(f'uploaded {input_path} ({row_count} row(s)) to {input_gcs_uri}') client = genai.Client( vertexai=True, project=settings.GOOGLE_CLOUD_PROJECT, location=location, ) start = datetime.now() start_monotonic = time.monotonic() job = await client.aio.batches.create( model=model, src=input_gcs_uri, config=types.CreateBatchJobConfig(dest=output_gcs_uri), ) print(f'submitted batch job: {job.name} (state={job.state})') await _poll_and_report( client.aio.batches, job, model=model, input_path=input_path, row_count=row_count, start=start, start_monotonic=start_monotonic, ) async def run_gemini_batch_job(model: str) -> None: api_key = os.environ.get('GEMINI_API_KEY') if not api_key: raise SystemExit( 'GEMINI_API_KEY is not set -- add a Gemini Developer API key ' 'to .env to use --api gemini. Never hardcode it here.' ) input_path = BATCH_INPUT_PATHS[model] row_count = _count_rows(input_path) client = genai.Client(vertexai=False, api_key=api_key) uploaded = await client.aio.files.upload( file=input_path, config=types.UploadFileConfig( display_name=input_path.stem, mime_type='jsonl' ), ) print(f'uploaded {input_path} ({row_count} row(s)) as file: {uploaded.name}') start = datetime.now() start_monotonic = time.monotonic() job = await client.aio.batches.create_embeddings( model=model, src=types.EmbeddingsBatchJobSource(file_name=uploaded.name), ) print(f'submitted batch job: {job.name} (state={job.state})') await _poll_and_report( client.aio.batches, job, model=model, input_path=input_path, row_count=row_count, start=start, start_monotonic=start_monotonic, ) def main() -> None: parser = argparse.ArgumentParser() parser.add_argument( '--model', required=True, choices=list(BATCH_INPUT_PATHS), help=( 'Which embedding model to submit the batch job for -- picks ' 'the matching batching_input_gemini_embedding_*.jsonl file, ' 'which this script only uploads, never generates.' ), ) parser.add_argument( '--api', default='vertex', choices=['vertex', 'gemini'], help=( "'vertex' (default): batches.create() via this project's " "normal Vertex credentials. 'gemini': batches.create_embeddings() " 'via the Gemini Developer API instead (requires GEMINI_API_KEY ' 'in .env) -- confirmed live that Vertex rejects ' 'create_embeddings() outright (ValueError: ... does not ' 'support batches.create_embeddings), so this is the only way ' 'to use that method. --location is ignored for --api gemini.' ), ) parser.add_argument( '--location', default=settings.GOOGLE_CLOUD_EMBEDDING_LOCATION, help=( 'Vertex AI location for the batch job -- only used for ' '--api vertex (default: settings.GOOGLE_CLOUD_EMBEDDING_LOCATION, ' f'currently {settings.GOOGLE_CLOUD_EMBEDDING_LOCATION!r}).' ), ) args = parser.parse_args() if args.api == 'vertex': asyncio.run(run_vertex_batch_job(args.model, args.location)) else: asyncio.run(run_gemini_batch_job(args.model)) if __name__ == ' __main__': main() ``` here my .jsonl file “batching\_input\_gemini\_embedding\_001.jsonl”, ```json { "key": "id_1", "request": { "content": { "parts": [ { "text": "dummy embedding text number 0" } ] }, "embed_content_config": { "output_dimensionality": 768, "title": "some_title_1", "task_type": "RETRIEVAL_DOCUMENT" } } } ``` run using ```bash uv run python assets/scripts/gemini/batching_embedding.py --model gemini-embedding-001 --api vertex --location asia-southeast1 ``` Also I tried inside the cookbook, [cookbook/quickstarts/Batch\_mode.ipynb at main · google-gemini/cookbook · GitHub](https://github.com/google-gemini/cookbook/blob/main/quickstarts/Batch_mode.ipynb) here my .jsonl file “batching\_input\_gemini\_embedding\_001.jsonl”, ```json { "key": "request_1", "request": { "output_dimensionality": 3, "content": { "parts": [ { "text": "1" } ] } } } ``` it outputted ```json { "key": "request_1", "request": { "content": { "parts": [ { "text": 1 } ] }, "output_dimensionality": 3 }, "status": "Bad Request: {\"error\": {\"code\": 400, \"message\": \"Invalid JSON payload received. Unknown name \\\"content\\\": Cannot find field.\\nInvalid JSON payload received. Unknown name \\\"output_dimensionality\\\": Cannot find field.\", \"status\": \"INVALID_ARGUMENT\", \"details\": [{\"@type\": \"type.googleapis.com/google.rpc.BadRequest\", \"fieldViolations\": [{\"description\": \"Invalid JSON payload received. Unknown name \\\"content\\\": Cannot find field.\"}, {\"description\": \"Invalid JSON payload received. Unknown name \\\"output_dimensionality\\\": Cannot find field.\"}]}]}}", "response": {}, "processed_time": "2026-09-20T02:34:53.662+00:00" } ``` I run on my machine using - python 3.12.11 - google-genai 2.22.0 NOTE 1. When I tried the payload for gemini-embedding-2 is work fine (just copy pasted what inside the documentation) What happen? Thank you