A WireMock simulation of Google Gemini's REST API — covering content generation, embeddings, files, cached contents, semantic retrieval, and tuned models.
A shared instance of this template runs on a public endpoint — call it right now from curl, an SDK, or an AI agent. Responses are fixed examples and behavior can't be changed; rate limits apply.
Jump to the endpoints ↓Run this template as a mock API in WireMock Cloud for a private instance you control — edit stubs, add stateful scenarios, use dynamic response templating, and inspect requests in the log.
Open in WireMock Cloud (free) →Live sandbox · no signup, no API key
A running WireMock Cloud instance of this template, callable right now from a terminal, a script, or an AI agent.
Endpoints
POST Generate Content with Tuned Model
curl -s -X POST 'https://8131q.wiremockapi.cloud/v1beta/tunedModels/your-tunedModelId:generateContent' \
-H 'Content-Type: application/json' \
-d '{}'POST Generate Text with Tuned Model
curl -s -X POST 'https://8131q.wiremockapi.cloud/v1beta/tunedModels/your-tunedModelId:generateText' \
-H 'Content-Type: application/json' \
-d '{}'POST Transfer Tuned Model Ownership
curl -s -X POST 'https://8131q.wiremockapi.cloud/v1beta/tunedModels/your-tunedModelId:transferOwnership' \
-H 'Content-Type: application/json' \
-d '{}'Spec catalog — Google Gemini API (v1beta) · 49 endpoints
| Method | Path | Summary |
|---|---|---|
| GET | /v1beta/cachedContents | List cached contents |
| POST | /v1beta/cachedContents | Create cached content |
| DELETE | /v1beta/cachedContents/{cachedContentId} | Delete cached content |
| GET | /v1beta/cachedContents/{cachedContentId} | Get cached content |
| PATCH | /v1beta/cachedContents/{cachedContentId} | Update cached content |
| GET | /v1beta/corpora | List corpora |
| POST | /v1beta/corpora | Create corpus |
| DELETE | /v1beta/corpora/{corpusId} | Delete corpus |
| GET | /v1beta/corpora/{corpusId} | Get corpus |
| GET | /v1beta/corpora/{corpusId}/permissions | List corpus permissions |
| POST | /v1beta/corpora/{corpusId}/permissions | Create corpus permission |
| DELETE | /v1beta/corpora/{corpusId}/permissions/{permissionId} | Delete corpus permission |
| GET | /v1beta/corpora/{corpusId}/permissions/{permissionId} | Get corpus permission |
| PATCH | /v1beta/corpora/{corpusId}/permissions/{permissionId} | Update corpus permission |
| GET | /v1beta/files | List files |
| POST | /v1beta/files | Upload a file |
| DELETE | /v1beta/files/{fileId} | Delete file |
| GET | /v1beta/files/{fileId} | Get file |
| GET | /v1beta/fileSearchStores | List file search stores |
| POST | /v1beta/fileSearchStores | Create file search store |
| DELETE | /v1beta/fileSearchStores/{fileSearchStoreId} | Delete file search store |
| GET | /v1beta/fileSearchStores/{fileSearchStoreId} | Get file search store |
| GET | /v1beta/models | List models |
| GET | /v1beta/models/{model} | Get model |
| POST | /v1beta/models/{model}:batchEmbedContents | Batch embed contents |
| POST | /v1beta/models/{model}:batchEmbedText | Batch embed text (PaLM) |
| POST | /v1beta/models/{model}:countMessageTokens | Count message tokens (PaLM) |
| POST | /v1beta/models/{model}:countTextTokens | Count text tokens (PaLM) |
| POST | /v1beta/models/{model}:countTokens | Count tokens |
| POST | /v1beta/models/{model}:embedContent | Embed content |
| POST | /v1beta/models/{model}:embedText | Embed text (PaLM) |
| POST | /v1beta/models/{model}:generateAnswer | Generate grounded answer |
| POST | /v1beta/models/{model}:generateContent | Generate content |
| POST | /v1beta/models/{model}:generateMessage | Generate message (PaLM) |
| POST | /v1beta/models/{model}:generateText | Generate text (PaLM) |
| POST | /v1beta/models/{model}:streamGenerateContent | Stream generate content |
| GET | /v1beta/tunedModels | List tuned models |
| POST | /v1beta/tunedModels | Create tuned model |
| DELETE | /v1beta/tunedModels/{tunedModelId} | Delete tuned model |
| GET | /v1beta/tunedModels/{tunedModelId} | Get tuned model |
| PATCH | /v1beta/tunedModels/{tunedModelId} | Update tuned model |
| POST | /v1beta/tunedModels/{tunedModelId}:generateContent | Generate content with tuned model |
| POST | /v1beta/tunedModels/{tunedModelId}:generateText | Generate text with tuned model (PaLM) |
| POST | /v1beta/tunedModels/{tunedModelId}:transferOwnership | Transfer tuned model ownership |
| GET | /v1beta/tunedModels/{tunedModelId}/permissions | List tuned model permissions |
| POST | /v1beta/tunedModels/{tunedModelId}/permissions | Create tuned model permission |
| DELETE | /v1beta/tunedModels/{tunedModelId}/permissions/{permissionId} | Delete tuned model permission |
| GET | /v1beta/tunedModels/{tunedModelId}/permissions/{permissionId} | Get tuned model permission |
| PATCH | /v1beta/tunedModels/{tunedModelId}/permissions/{permissionId} | Update tuned model permission |
Unauthenticated WireMock Cloud demo sandbox for Google Gemini — not an official Google Gemini sandbox, and it returns simulated example data only.
Generated from the v1beta Gemini API description, the template answers models/{model}:generateContent the way the real service does — a candidates array whose content carries the reply parts, per-candidate safety ratings, and a usageMetadata block counting prompt and candidate tokens. Code that unwraps real Gemini responses, including the google-genai SDK's own parsing, runs against it unchanged.
Google's google-genai SDK routes every call through an http_options base URL, so one constructor field moves a client from the live service to the mock:
Frameworks that wrap the SDK — agent runtimes, LangChain-style toolchains — usually surface that same base-URL setting, so they can be aimed at the sandbox, or at your own copy of the mock, without forking any code.
A fixed Gemini backend turns flaky problems into solved ones: CI that needs no Google credentials, prompt-pipeline tests that can assert on exact output, and cost-free load on the request path while you profile your own service.
from google import genai
from google.genai.types import HttpOptions
# any api_key satisfies the client —
# the mock ignores it
client = genai.Client(
api_key="any-value",
http_options=HttpOptions(base_url="https://8131q.wiremockapi.cloud"),
)
reply = client.models.generate_content(
model="gemini-2.0-flash",
contents="Explain how AI works",
)
print(reply.text)Gemini clients send the key as a query parameter or an x-goog-api-key header; the mock accepts anything in either place and inspects neither. A genuine Google Cloud key should still stay out of the picture — the sandbox is public infrastructure with no auth in front of it.
The route exists, but it answers with a single JSON body instead of chunked partial candidates. Verify request wiring against it, and move to a private WireMock Cloud instance if your assertions depend on genuine chunk-by-chunk delivery.
The v1beta surface Google publishes still defines the legacy generateText, generateMessage, and embedText operations, so the template stubs them too — useful when you're migrating an older PaLM integration and want both API generations answering side by side.
Not from the shared instance, which sticks to successful generations. Recreate the template under your own account and stub the failure shapes you need — a candidate with finishReason SAFETY, or a 429 carrying RESOURCE_EXHAUSTED — and drive your fallback logic through them deterministically.
Vendor names identify APIs represented by WireMock template sources. This page does not imply vendor endorsement, certification, partnership, or official integration status.