mwcyu unifies Mike's chat engine and hands every model web search

A single streaming core replaces three hand-built ones, and once it landed, giving every AI model live web search became almost free.

searchinfrastructure

mwcyu opened a branch to consolidate how Mike streams answers back to the user, one word at a time. Instead of each underlying AI model getting its own hand-wired streaming code, the work pulls everything behind one shared engine - the kind of plumbing that pays off later. What it unlocked:

One streaming core, so every AI model returns answers through the same tested path rather than three separate implementations.
Live web search on all three models, letting the assistant pull in current information mid-answer instead of relying only on what it was trained on.
A backend test layer covering all of Mike's API routes, the largest single piece of the work and a safety net around the rewiring.

One catch: the pull request was closed without merging, so none of this reached the main Mike project.

So what Worth a look for anyone weighing how to add live web search to a legal assistant without maintaining separate code per AI provider.

View this fork on GitHub →

Spotted something wrong? Or know the PR text has fresher detail than the writeup above?

Commits in this thread

4 commits from mwcyu/Mike-fork, oldest first. Source extracted verbatim from the harvested git log.

SHA	Subject	Author	Date
`e688ccbc`	Use OpenAI SDK for Responses API streaming	Claude	2026-05-31	↗ GitHub
commit body Replace the hand-rolled fetch + SSE parsing in the OpenAI adapter with the official openai package. client.responses.create({ stream: true }) returns a typed async-iterable of events, eliminating the manual TextDecoder/buffer and extractSseJson machinery. Behavior, exports, and the tool-call loop are unchanged.
`9dfb6fd5`	Extract shared streaming loop into a provider-agnostic driver	Claude	2026-05-31	↗ GitHub
commit body The Claude, OpenAI, and Gemini adapters each re-implemented the same agentic streaming loop (iterate to maxIterations, stream a turn, accumulate fullText, run tools, feed results back). That triplication let the loop logic drift between providers. Introduce backend/src/lib/llm/driver.ts owning the loop, break conditions, the runTools call, and the single fullText accumulation. Each provider becomes a thin session factory (create{Claude,OpenAI, Gemini}Session) that owns its SDK call, event parsing, follow-up message state, and all callback firing. Public stream* signatures are unchanged, so callers stay untouched. Behavior is preserved: per-provider callback ordering, the OpenAI pre-tool preamble drop, Claude's stop_reason hard-stop, Gemini's verbatim thoughtSignature replay, and OpenAI instructions-on-iter-0. Chat history persistence (route-level, fed by callbacks + fullText) is unaffected. Typecheck passes.
`0ad1b344`	Add native web-search tool to all three LLM adapters	Claude	2026-05-31	↗ GitHub
commit body Enable each provider's built-in, server-executed web search so the main chat can browse the internet: - Claude: native web_search_20250305 server tool - OpenAI Responses: web_search built-in tool - Gemini: googleSearch grounding tool These run server-side (the provider performs the search and folds results into its answer), so they bypass the runTools function-call loop. A new enableWebSearch flag on StreamChatParams gates them; only the interactive main chat (runLLMStream) opts in, leaving tabular review and bulk extraction untouched. Surface searches to the UI via a new onWebSearch callback that each adapter fires when a search starts - Claude on the server_tool_use contentBlock, OpenAI on the web_search_call output item, Gemini on groundingMetadata.webSearchQueries. runLLMStream streams a web_search event and persists it to the assistant message (unlike the transient tool_call_start), so reloaded chats still show "Searched the web for ...". Added the web_search variant to the backend and frontend AssistantEvent unions, a stream handler, and a WebSearchBlock renderer. Backend and frontend typecheck clean.
`a101800d`	test(backend): add Vitest route integration tests for all 8 routers	Claude	2026-06-01	↗ GitHub
commit body Introduce a supertest-based integration suite covering user, chat, projectChat, projects, documents, tabular, workflows and downloads routes (77 tests). Supabase, auth, storage and the LLM/chat tooling are mocked; streaming SSE endpoints and the signed-download token round-trip are exercised end to end. Refactor src/index.ts to export createApp() and only bind a port when run directly, and skip rate limiters under NODE_ENV=test so the suite is deterministic. Add vitest + supertest dev deps and test scripts.

Capture this thread into my fork

Download a single Markdown prompt that tells Claude how to port every commit above into your working tree — adapting paths and structure to match your repo. Run it via claude -p < capture-thread-747.md from inside the repo you want the changes in.

⬇ Download capture-thread-747.md