In past I managed to use variety of databases as an AI memory for the Hermes Agent. Like I tried SQLite, and few others. I decided to try the SQL Server recently.
Why SQL Server? It's SQL. And it can keep the NoSQL as well as the vector data. So we have best of both worlds benefits.
If you are using the Hermes agent then it helps to use the local databases like Postgres, SQL Server and similar other databases. So it makes sense for the databases to be used for storing the Hermes agent memories.
In recent demo I managed to work with SQL Server to successfully connect the local SQL server with the Hermes agent. You can check out the video below.
SQL Server 2025 now accepts the vector database support. And also this database can be used for variety of the hermes agent's localized memory layer. You can save this and also retrieve using SQL queries.
If you are planning on building anything with the SQL server then also getting used to your agent with this database makes things super easy and faster.
I wish to now cover most of the NoSQL based database. I do have yet to cover the MySQL database and also other databases like PostgreSQL and few others. Firebase was something I was meaning to add I would cover this in upcoming posts.
Have you tried adding SQL Server database as a memory layer in the Hermes Agent? Do let me know if you have tried it.
Top comments (0)