Is session memory unsupported for Ollama? (openai embeddings 401)
Hey, quick question about session memory + Ollama.
I’m running OpenClaw 2026.2.1 locally with
ollama/qwen2.5:7b-instruct-q4_K_M and Ollama detected fine for chat.
When session-memory is enabled, I consistently get this error:
It looks like memory embeddings are still routed through the OpenAI client, even when using Ollama as the model provider.
Question:
Is session memory currently unsupported with Ollama (local) in OpenClaw?
Or is there a documented way to route embeddings to Ollama / disable the OpenAI embeddings path?
I couldn’t find any mention of Ollama embeddings or local memory config in the docs, so I want to confirm if this is a known limitation vs a misconfig.
I’m running OpenClaw 2026.2.1 locally with
ollama/qwen2.5:7b-instruct-q4_K_M and Ollama detected fine for chat.
When session-memory is enabled, I consistently get this error:
It looks like memory embeddings are still routed through the OpenAI client, even when using Ollama as the model provider.
Question:
Is session memory currently unsupported with Ollama (local) in OpenClaw?
Or is there a documented way to route embeddings to Ollama / disable the OpenAI embeddings path?
I couldn’t find any mention of Ollama embeddings or local memory config in the docs, so I want to confirm if this is a known limitation vs a misconfig.
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
83,815Members
Resources
Recent Announcements
Similar Threads
Was this page helpful?
