kunthawat/moreminimore-vibe
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Replace thinking with native Gemini thinking summaries (#400)

Browse Source 
This uses Gemini's native [thinking
summaries](https://cloud.google.com/vertex-ai/generative-ai/docs/thinking#thought-summaries)
which were recently added to the API.

Why? The grafted thinking would sometimes cause weird issues where the
model, especially Gemini 2.5 Flash, got confused and put dyad tags like
`<dyad-write>` inside the `<think>` tags.

This also improves the UX because you can see the native thoughts rather
than having the Gemini response load for a while without any feedback.

I tried adding Anthropic extended thinking, however it requires temp to
be set at 1, which isn't ideal for Dyad's use case where we need precise
syntax following.
This commit is contained in:
Will Chen
2025-06-16 17:29:32 -07:00
committed by GitHub
parent 3a6ab12bed
commit 30b5c0d0ef
39 changed files with 1020 additions and 628 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
e2e-tests/engine.spec.ts
View File

@@ -11,15 +11,29 @@ testSkipIfWindows("send message to engine", async ({ po }) => {
await po.snapshotMessages({ replaceDumpPath: true });
});
testSkipIfWindows("send message to gateway", async ({ po }) => {
testSkipIfWindows("send message to engine - openai gpt-4.1", async ({ po }) => {
await po.setUpDyadPro();
await po.selectModel({ provider: "Google", model: "Gemini 2.5 Flash" });
await po.sendPrompt("[dump] tc=gateway-simple");
// By default, it's using auto which points to Flash 2.5 and doesn't
// use engine.
await po.selectModel({ provider: "OpenAI", model: "GPT 4.1" });
await po.sendPrompt("[dump] tc=turbo-edits");
await po.snapshotServerDump("request");
await po.snapshotMessages({ replaceDumpPath: true });
});
testSkipIfWindows(
"send message to engine - anthropic claude sonnet 4",
async ({ po }) => {
await po.setUpDyadPro();
// By default, it's using auto which points to Flash 2.5 and doesn't
// use engine.
await po.selectModel({ provider: "Anthropic", model: "Claude 4 Sonnet" });
await po.sendPrompt("[dump] tc=turbo-edits");
await po.snapshotServerDump("request");
},
);
// auto (defaults to Gemini 2.5 Flash)
testSkipIfWindows("auto should send message to gateway", async ({ po }) => {
await po.setUpDyadPro();