[flight-booking-agent] Multi-turn chat and message injection (variant A) #25
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Peter Wielander <[email protected]>
Signed-off-by: Peter Wielander <[email protected]>
Signed-off-by: Peter Wielander <[email protected]>
Signed-off-by: Peter Wielander <[email protected]>
Signed-off-by: Peter Wielander <[email protected]>
Signed-off-by: Peter Wielander <[email protected]>
VaguelySerious
changed the title
![vercel[bot]](https://avatars.githubusercontent.com/in/8329?s=60&v=4){kind=small}
Signed-off-by: Peter Wielander <[email protected]>
…g added to conversation history in multi-turn conversations Co-authored-by: VaguelySerious <[email protected]>
Signed-off-by: Peter Wielander <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This makes multi-turn session the default for the flight booking app. This is an alternative to #26, using a different pattern for user message persistence.
The docs PR vercel/workflow#759 also publishes the page on how to do chat session modeling, which includes example code snippets from this PR.
Critically, this PR also removes all
localStorageuse for persistence, with the exception of the last run ID. The durable streams serve as the main durability and storage. To do this, we also specifically persist user messages inside the stream, which wasn't done previouslyAlso see vercel/workflow#739
Difference in variants
Variant A lets
useChatnatively handle optimistic rendering of user messages, and writes custom data chunks to the stream that the client interprets as user messages.Variant B writes user messages to the actual stream as user messages, so that the stream is the source of truth and consists of 100% normally interpretable
UIMessageChunkchunks.I prefer variant B, as it's better for later processing with AI SDK, since the user messages are inside the stream as the SDK would expect them to be.
The downside with Variant B is that is requires slightly more optimistic rendering code.
Screenshots