- Who it helps
- Developers testing streaming, error handling and tool-call parsing.
- Availability
- Public Go repository checked October 29. Building requires Go 1.24 or later; the mock needs no model weights or GPU.
Responses a client can rehearse
The server can simulate streamed answers, tool calls, JSON and mid-stream failures. Its default address is 127.0.0.1:11434; choose a different port when real Ollama already uses it.
OllamaMock: exercise a model client without model inference ↗
What a mock test proves
A passing mock test can establish client behavior under the simulated response. It cannot establish answer quality or actual model speed.
OllamaMock: exercise a model client without model inference ↗OllamaMock maker overview ↗
What you can do
- Inspect the mock response modes and port configuration. Open reference ↗
Where the evidence stops
Documentation only. No server or third-party project was executed.
Sources read. Sources checked October 29, 2026. No installation, model inference or end-to-end workflow was executed.
Sources & further reading
GitHub · official
OllamaMock: exercise a model client without model inference ↗
Supports ollamaMock: exercise a model client without model inference.
kinduff · page checked 2026-10-29
Web · official
OllamaMock maker overview ↗
Supports ollamaMock maker overview.
kinduff · page checked 2026-10-29
Coverage note: No qualifying current independent attention sample; original X posts unavailable. Reported outcomes were not independently reproduced.
The record over time (1)
Dated source checks and published explanations. An observation date can differ from the original release date.
- · Added to Current
OllamaMock supplies predictable fake responses to clients that speak the Ollama API.