teasa.aiLorebook Lab
Edit world guidance, test keyword activation, and bring the result into a Teasa Plot.
Loading the local editor.
Find the keyword that pulls in the wrong setting
Give the Sunken Bell tavern the keyword ‘bell’, and a harbor warning can pull in tavern lore. Open the example, save a revision, replace ‘bell’ with ‘Sunken Bell’, and compare the same three messages.
The harbor is an original Teasa example. Results are calculated entry selections, not recorded AI conversations.
Choose ‘Harbor: repaired keyword’ in the examples menu to open the one-keyword repair with the original saved for comparison. Run the same three tests and inspect what changed. Replacing your workspace requires confirmation.
From a tested book to a playable Plot
- Edit titles, instructions, and keywords. Add recent messages to see where a match came from.
- Set the entries you expect for each input, then compare revisions.
- Choose Use in Teasa, sign in, and review the entries. Save a private Blueprint and connect it to a new or existing draft.
- Complete the Plot’s cast and opening, then use Play-test to inspect real responses before publishing.
What does this tool check?
The lab runs Teasa’s current entry-selection code against your input and the last four supplied messages, including entry-count and context-size limits. It does not emulate another service’s retrieval engine or inspect a live chat. Selecting a fact does not guarantee that an AI will use it correctly.
What does a shared link contain?
A workspace link contains every entry’s text. Test messages and history are included only if you choose them. Anyone with the link can read its contents. Ordinary editing and tests do not send your writing to a server.