Answer emails from your knowledge base
The outcome: someone asks what your return window is, what a plan costs, or whether you ship to Norway — and the reply comes back with your answer, taken from material you already have instead of invented.
There is more than one way to give atbridge.ai your own knowledge, and they behave differently. Start from what you already have.
Use an OpenAI Vector Store You already keep a product corpus, handbook or website content in OpenAI. Looked up automatically while it drafts.
Use your atbridge Notes Your knowledge is what you write and edit yourself. The AI browses, searches and reads your notes as it works.
Connect Obsidian to Thunderbird You already keep a Markdown vault. Point the Notes folder at it and everything on the Notes page applies.
Which one fits
Section titled “Which one fits”The important difference is not the storage — it is when the knowledge reaches the model.
| OpenAI Vector Store | atbridge Notes (and Obsidian) | |
|---|---|---|
| Best when | the material is already a corpus you sync | you write and edit the material yourself |
| How the AI gets it | looked up automatically while it drafts | the AI searches and reads notes as a tool, or you hand one over |
| While writing in the compose dock | works | not available — the dock has no notes tools |
| Providers | OpenAI only | any provider that can use tools |
| Plan | Pro (it rides on Connect) | Personal or Pro |
| Keeping it current | you sync the store | you edit the notes |
Where to go next
Section titled “Where to go next”- Use an OpenAI Vector Store — set up in one field, then it just happens
- Use your atbridge Notes — no setup at all if you already keep notes
- Connect Obsidian to Thunderbird — one folder setting, then read the Notes page
Was this page helpful?
Thanks for the feedback!