Docs
One Click Deploy You Own Hybrid AI Search Engine
One Click Deploy You Own Hybrid AI Search Engine
How to one click deploy MemFree
What could you do with Open Source MemFree
- You could deploy your own Hybrid AI Search Engine bases on your bookmarks, notes, and docs
- You can start by indexing the best resources on a topic and then build a domain-specific AI search engine
- You can first index all the documents of a website, and then build an intelligent answer assistant for a website
How to One Click Deploy
Deploy Backend with Zeabur
- You could get
UPSTASH_REDIS_REST_URL
andUPSTASH_REDIS_REST_TOKEN
from Upstash Redis - You could use
openssl rand -base64 32
to generate a random string for API Token
Deploy Frontend with Vercel
- You could get
OPENAI_API_KEY
from OpenAI - The MEMFREE_HOST is
https://${PUBLIC_DOMAIN}.zeabur.app
- The AUTH_SECRET is a random string, you could use
openssl rand -base64 32
to generate it - The API_TOKEN should be the same as the one you used in the backend
Some Optional variables
- To enable faster search, You could set
SERPER_API_KEY
in Vercel, you could getSERPER_API_KEY
from Serper Dev - To enable AXIOM log, you could set
AXIOM_TOKEN
in Vercel and Zeabur, you could getAXIOM_TOKEN
from Axiom - To enable faster web page parser, you could set
JINA_KEY
in Zeabur, you could getJINA_KEY
from Jina Reader
3.1 AWS S3 Express Bucket
In order to facilitate developers to quickly deploy with one click, memfree's vector database will store data in a temporary local directory. If you want to deploy in a production environment, please use AWS S3 Express Bucket
- You could get AWS variables Refer to How To Create S3 Express Bucket
- You need to set the following variables in Zeabur Vector Service
AWS_ACCESS_KEY_ID
AWS_SECRET_ACCESS_KEY
AWS_REGION
AWS_BUCKET