mem0 setup for long term memory
✅Solved
Lets say I want to setup mem0 with openclaw while using local embedding model such as Qwen3-Embedding-4b / nomic-embed-text + chroma
Do i need to modify the SOUL.md to tell openclaw to use the new feature such as :
memory_search
memory_list
memory_store
memory_get
memory_forget
instead of the default storing in memory.md ? Or I just have to setup the plugin and the openclaw will figure out the rest ?
https://mem0.ai/blog/add-persistent-memory-openclaw#memory-is-captured-automatically
Do i need to modify the SOUL.md to tell openclaw to use the new feature such as :
memory_search
memory_list
memory_store
memory_get
memory_forget
instead of the default storing in memory.md ? Or I just have to setup the plugin and the openclaw will figure out the rest ?
https://mem0.ai/blog/add-persistent-memory-openclaw#memory-is-captured-automatically
Solution
Nice — if
The reason OpenClaw/Mem0 is still throwing “Incorrect API key” is a Mem0 OSS TypeScript limitation:
Mem0 OSS
In the current TS OSS SDK, the OpenAI embedder is literally:
(no
That means your vectorengine key gets sent to OpenAI → OpenAI rejects it → 401.
So you’ve got 3 workable options:
curlhttps://api.vectorengine.ai/v1/embeddingsAuthorization: Bearer …The reason OpenClaw/Mem0 is still throwing “Incorrect API key” is a Mem0 OSS TypeScript limitation:
Mem0 OSS openai
embedder ignores baseURL
In the current TS OSS SDK, the OpenAI embedder is literally:openaibaseURL(no
baseURLhttps://api.openai.com/v1baseURLThat means your vectorengine key gets sent to OpenAI → OpenAI rejects it → 401.
So you’ve got 3 workable options:
Option A — Use a real OpenAI API key for embeddings
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
134,827Members
Resources
Recent Announcements
Similar Threads
Was this page helpful?
