Harden score api and fern rerun - #29
Open
kgowru wants to merge 10 commits into
Open
Conversation
Replace gpt-4o-mini with claude-haiku-4-5 in the LLM fallback for category classification. Uses ANTHROPIC_API_KEY instead of OPENAI_API_KEY. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
feat: switch category inference from OpenAI to Claude Haiku
Add a slug-alias map so typing a domain like monday.com (which resolves to the weaker "monday" slug) lands on the curated leaderboard entry "developer-monday-com-apps". Applied in the score API and the company page so both search and direct navigation redirect to the canonical slug. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Redirect duplicate company slugs to canonical leaderboard entry
) Fern preview/staging hosts (*.ferndocs.com) now always slug by URL instead of the derived company name, so e.g. docusign.ferndocs.com no longer collapses onto the "docusign" slug and return the cached live entry. These sites are now graded as distinct entries. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix: route monday.com to canonical api-reference leaderboard entry The "monday" alias pointed at developer-monday-com-apps, which has since been hidden (score 68/D). Repoint to developer-monday-com-api-reference, the visible canonical entry (score 91/A). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix: make slug alias a lookup redirect, never a scoring target Previously a forced rerun (or dev run) of monday.com would score https://monday.com and upsert it under the aliased slug developer-monday-com-api-reference, clobbering the curated A-grade entry. Now the alias only redirects cache lookups to the existing canonical entry; scoring always stores under the raw slug. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…om) (#27) An llms.txt returning 200 short-circuited docs detection, so marketing homepages that now ship one (monday.com) were graded as docs sites. At a bare apex/root, llms.txt is no longer sufficient on its own; we defer to the homepage content check (only a visible marketing page is rejected, and deeper paths / docs subdomains still pass). Also drop the now-unneeded monday slug alias. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The detection layer already rejects monday.com, but the cache lookup runs before detection — a stale cached row would still be served. Add an apex-only block (exact host + www, subdomains excluded) that runs before the cache and scoring, guaranteeing the marketing apex is rejected while developers.monday.com and other docs subdomains stay gradeable. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Replace the grade-grouped grid and filter tabs with a grid of category panels: each panel shows ranked rows with favicon, name, colored score and grade, top 5 per category with a View all toggle. Responsive at 3/2/1 columns. Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
#38) * fix: use VERCEL_ENV to gate production assetPrefix NODE_ENV is 'production' for both production and preview builds on Vercel, so the absolute assetPrefix was forced onto preview deployments too — their HTML then loaded CSS/JS from the production domain with mismatched chunk hashes, rendering previews unstyled. Keying off VERCEL_ENV keeps the absolute prefix on real production (needed because the site is proxied under buildwithfern.com/agent-score) while previews serve assets from their own URL. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix scoring error text wrap and swap hero testimonial logos Let the "Last checked … · Rerun …" line wrap instead of forcing nowrap, so the scoring timeout error no longer overflows into the simulation panel. Also swap the two hero quote cards from OpenRouter/Resend to ElevenLabs/Merge. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
No description provided.