feat: split data update from site build; publish via Cloudflare Pages

The updater fetched GitHub metadata and rendered the dashboard payload in
a single step, so anything that published the site needed a GITHUB_TOKEN.
Cloudflare Pages builds from a Git webhook and has no business holding
one.

Split the tool into three modes. The default update step still fetches
GitHub and now records every API-sourced field in data/metadata.json,
which is committed alongside README.md and data/history.jsonl. The new
-build mode joins that snapshot with data/agents.yml and
data/history.jsonl to render dist/ with no network access and no token;
-check is unchanged.

Supporting changes:

- computeDeltaAt anchors the delta windows to when the data was fetched
  rather than the wall clock, so a redeploy days later reproduces the
  same Δ7d instead of sliding the window past its slack allowance.
- writeSiteData takes updatedAt explicitly for the same reason: the
  timestamp labels data freshness, not build time.
- sortStats is extracted from fetchStats so the build step re-ranks
  identically from committed metadata.
- An agents.yml entry with no metadata yet is omitted with a warning
  instead of failing the build, which would otherwise block every deploy
  between merging a new entry and the next nightly run.
- update.yml drops the GitHub Pages deploy steps and commits
  data/metadata.json; ci.yml runs `go run . -build` so a build that would
  break on deploy breaks in CI first.
- site/_headers stops the edge serving a stale data.json after a refresh.
- docs/DEPLOY.md covers the Cloudflare setup, including the one-time
  bootstrap of data/metadata.json that the build depends on.

Also carries the in-progress curation work already in the tree: the
module rename to awesome-ai-dev-tools, retagged entries, and removal of
the archived Roo-Code, void, continue and suna entries.

.github/workflows/ci.yml

@@ -11,6 +11,7 @@ on:
       - '.github/workflows/ci.yml'
       - 'data/agents.yml'
       - 'templates/**'
+      - 'site/**'
 
 permissions:
   contents: read
@@ -30,6 +31,9 @@ jobs:
       - run: go test ./...
       - run: go build ./...
       - run: go run . -check
+      # Runs exactly what Cloudflare Pages runs, so a build that would fail on
+      # deploy fails here instead. Needs no token — that is the point of the split.
+      - run: go run . -build
 
   lint:
     runs-on: ubuntu-latest

.github/workflows/update.yml

@@ -8,7 +8,6 @@ on:
     branches: [main]
     paths:
       - 'data/agents.yml'
-      - 'site/**'
       - 'templates/**'
       - '**.go'
       - 'go.mod'
@@ -18,11 +17,11 @@ permissions:
   # Uses GITHUB_TOKEN intentionally — a PAT would cause infinite trigger loops
   # on the auto-commit ("chore: daily ranking refresh") because GITHUB_TOKEN-authored
   # pushes do NOT re-trigger workflows, whereas a PAT would.
+  #
+  # This job only refreshes data. Publishing is Cloudflare Pages' job: it builds
+  # from the commit this job pushes (webhook-driven, so the GITHUB_TOKEN caveat
+  # above does not apply to it) by running `go run . -build`, which needs no token.
   contents: write
-  # Pages deploy happens in THIS workflow (a separate push-triggered Pages
-  # workflow would never fire — see GITHUB_TOKEN note above).
-  pages: write
-  id-token: write
 
 concurrency:
   group: update-rankings
@@ -31,9 +30,6 @@ concurrency:
 jobs:
   update:
     runs-on: ubuntu-latest
-    environment:
-      name: github-pages
-      url: ${{ steps.deploy.outputs.page_url }}
     steps:
       - uses: actions/checkout@v7
 
@@ -42,6 +38,9 @@ jobs:
           go-version: 'stable'
           cache: true
 
+      # The only step that touches the network or needs a token. It refreshes
+      # README.md, data/history.jsonl and data/metadata.json; rendering the
+      # site from those files happens later, on Cloudflare.
       - name: Run updater
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
@@ -51,7 +50,7 @@ jobs:
         run: |
           git config user.name  "github-actions[bot]"
           git config user.email "41898282+github-actions[bot]@users.noreply.github.com"
-          git add README.md data/history.jsonl
+          git add README.md data/history.jsonl data/metadata.json
           if git diff --staged --quiet; then
             echo "no changes"
             exit 0
@@ -59,7 +58,7 @@ jobs:
           git commit -m "chore: daily ranking refresh"
           # Rebase + retry guards against the race where two runs commit near-simultaneously.
           # Strategy: plain rebase first; only force-resolve bot-generated files (README.md,
-          # data/history.jsonl) when they are the sole conflicts. If any human-maintained file
+          # data/history.jsonl, data/metadata.json) when they are the sole conflicts. If any human-maintained file
           # (e.g. data/agents.yml) conflicts, abort and fail loudly so a human can investigate.
           for attempt in 1 2 3; do
             if git push; then exit 0; fi
@@ -72,31 +71,15 @@ jobs:
             conflicted=$(git diff --name-only --diff-filter=U)
             echo "conflicted files: $conflicted"
             # Only auto-resolve if ALL conflicts are in bot-generated files.
-            non_bot=$(echo "$conflicted" | grep -v -E '^(README\.md|data/history\.jsonl)$' || true)
+            non_bot=$(echo "$conflicted" | grep -v -E '^(README\.md|data/history\.jsonl|data/metadata\.json)$' || true)
             if [ -n "$non_bot" ]; then
               echo "ERROR: conflict in human-maintained file(s): $non_bot — aborting rebase"
               git rebase --abort
               exit 1
             fi
             # Safe to force-resolve: keep our (freshest) bot-generated content.
-            git checkout --theirs README.md data/history.jsonl 2>/dev/null || true
+            git checkout --theirs README.md data/history.jsonl data/metadata.json 2>/dev/null || true
-            git add README.md data/history.jsonl
+            git add README.md data/history.jsonl data/metadata.json
             GIT_EDITOR=true git rebase --continue
           done
           exit 1
-
-      # site/data.json was generated by the updater run above; deploy the
-      # static dashboard (site/) to GitHub Pages every run.
-      - name: Configure Pages
-        uses: actions/configure-pages@v6
-        with:
-          enablement: true
-
-      - name: Upload Pages artifact
-        uses: actions/upload-pages-artifact@v5
-        with:
-          path: site
-
-      - name: Deploy to GitHub Pages
-        id: deploy
-        uses: actions/deploy-pages@v5

.gitignore

@@ -1,5 +1,5 @@
 # Go build artifacts
-/ai-dev-tools
+/awesome-ai-dev-tools
 *.exe
 *.test
 *.out
@@ -9,5 +9,7 @@
 .idea/
 .vscode/
 
-# Generated by the updater each run; deployed to Pages, never committed
+# Build output from `go run . -build`; regenerated on every deploy.
-site/data.json
+# data/metadata.json IS committed — it is the update step's handoff to the
+# build step, and the reason the build needs no GITHUB_TOKEN.
+/dist/

README.md

@@ -1,9 +1,9 @@
-# AI Dev Tools
+# Awesome AI Dev Tools
 
 > Curated ranking of AI-powered developer tools, sorted by GitHub stars.
 > Updated daily by GitHub Actions.
 
-📊 **[Interactive dashboard with star-history charts →](https://tiennm99.github.io/ai-dev-tools/)**
+📊 **[Interactive dashboard with star-history charts →](https://awesome-ai-dev-tools.pages.dev/)**
 
 **Last updated:** 2026-09-16 03:54 UTC · **Tracked:** 40 repos
 **Top 7-day mover:** [earendil-works/pi](https://github.com/earendil-works/pi) (+2642 stars)
@@ -59,8 +59,8 @@
 2. A daily GitHub Actions workflow (`.github/workflows/update.yml`) runs the Go updater.
 3. The updater fetches live repo metadata via the GitHub GraphQL API in one batched query.
 4. Star counts are appended to `data/history.jsonl` for 7-day delta computation.
-5. This `README.md` is regenerated from `templates/readme.tmpl` and committed back to the repo.
+5. This `README.md` and `data/metadata.json` are regenerated and committed back to the repo.
-6. `site/data.json` is regenerated and the [dashboard](https://tiennm99.github.io/ai-dev-tools/) (`site/index.html`) is redeployed to GitHub Pages.
+6. That push triggers Cloudflare Pages, which runs `go run . -build` to render the [dashboard](https://awesome-ai-dev-tools.pages.dev/) from the committed data — no API token needed at build time. See [docs/DEPLOY.md](./docs/DEPLOY.md).
 
 **Δ7d:** Change in stars over the past 7 days; `—` means fewer than 7 days of history.
 
@@ -83,7 +83,7 @@ agents:
     tags: [terminal, byo-model, interactive, community]
 ```
 
-**Tags** describe a tool across five facets — the [dashboard](https://tiennm99.github.io/ai-dev-tools/) filters on them, OR within a facet and AND across facets:
+**Tags** describe a tool across five facets — the [dashboard](https://awesome-ai-dev-tools.pages.dev/) filters on them, OR within a facet and AND across facets:
 
 | Facet | Tags |
 |-------|------|

build.go

@@ -0,0 +1,108 @@
+package main
+
+import (
+	"fmt"
+	"io"
+	"os"
+	"path/filepath"
+)
+
+// runBuild renders the static dashboard into distDir from committed inputs
+// only — data/agents.yml, data/metadata.json and data/history.jsonl. It makes
+// no network calls and needs no GITHUB_TOKEN, which is what lets an untrusted
+// build environment (Cloudflare Pages) run it.
+//
+// Splitting this out from the update step also means a pure curation change
+// (retagging an entry, editing a note) republishes immediately on push,
+// reusing the last fetched star figures instead of waiting for the nightly run.
+func runBuild(agentsPath, metadataPath, historyPath, siteDir, distDir string) error {
+	agents, err := loadAgents(agentsPath)
+	if err != nil {
+		return err
+	}
+
+	meta, err := readMetadata(metadataPath)
+	if err != nil {
+		return err
+	}
+
+	stats := statsFromMetadata(agents, meta)
+	stats = enforceStarFloor(stats)
+	if len(stats) == 0 {
+		return fmt.Errorf("no entries left to publish after joining %s with %s", agentsPath, metadataPath)
+	}
+	sortStats(stats)
+
+	history, err := readSnapshots(historyPath)
+	if err != nil {
+		return err
+	}
+
+	// Anchor the delta windows to when the data was fetched, not to wall clock.
+	// A rebuild triggered days later (a docs push, a manual redeploy) must show
+	// the same "Δ7d" the nightly run computed, not a window silently slid
+	// forward past its slack allowance.
+	current := Snapshot{
+		Date:  meta.FetchedAt.UTC().Format("2006-01-02"),
+		Stars: make(map[string]int, len(stats)),
+	}
+	for _, s := range stats {
+		current.Stars[s.CanonicalKey] = s.Stars
+	}
+	deltas7 := computeDeltaAt(history, current, meta.FetchedAt, 7, 3)
+	deltas30 := computeDeltaAt(history, current, meta.FetchedAt, 30, 5)
+
+	if err := copyDir(siteDir, distDir); err != nil {
+		return fmt.Errorf("copy %s to %s: %w", siteDir, distDir, err)
+	}
+
+	updatedAt := meta.FetchedAt.UTC().Format("2006-01-02 15:04 UTC")
+	if err := writeSiteData(filepath.Join(distDir, "data.json"), updatedAt, stats, deltas7, deltas30, history); err != nil {
+		return err
+	}
+
+	fmt.Printf("built %s: %d entries, data fetched %s\n", distDir, len(stats), updatedAt)
+	return nil
+}
+
+// copyDir replaces dst with a fresh copy of src. The wipe matters: dist/ is
+// build output, and a file deleted from site/ must not survive in a deploy
+// because a previous build left it there.
+func copyDir(src, dst string) error {
+	if err := os.RemoveAll(dst); err != nil {
+		return err
+	}
+	return filepath.WalkDir(src, func(path string, d os.DirEntry, err error) error {
+		if err != nil {
+			return err
+		}
+		rel, err := filepath.Rel(src, path)
+		if err != nil {
+			return err
+		}
+		target := filepath.Join(dst, rel)
+		if d.IsDir() {
+			return os.MkdirAll(target, 0o755)
+		}
+		if !d.Type().IsRegular() {
+			return nil // skip symlinks and other irregular entries
+		}
+		return copyFile(path, target)
+	})
+}
+
+func copyFile(src, dst string) error {
+	in, err := os.Open(src)
+	if err != nil {
+		return err
+	}
+	defer func() { _ = in.Close() }() // read-only; close error is not actionable
+
+	if err := os.MkdirAll(filepath.Dir(dst), 0o755); err != nil {
+		return err
+	}
+	return atomicWriteFile(dst, func(w io.Writer) error {
+		_, err := io.Copy(w, in)
+		return err
+	})
+}

build_test.go

@@ -0,0 +1,239 @@
+package main
+
+import (
+	"encoding/json"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+)
+
+// writeBuildFixture lays down a minimal repo layout the build step can read:
+// agents.yml (curation), metadata.json (fetched figures), history.jsonl
+// (delta baselines) and a site/ source directory.
+func writeBuildFixture(t *testing.T, fetchedAt time.Time) (dir string) {
+	t.Helper()
+	dir = t.TempDir()
+
+	mustWrite := func(rel, content string) {
+		t.Helper()
+		path := filepath.Join(dir, rel)
+		if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
+			t.Fatalf("mkdir for %s: %v", rel, err)
+		}
+		if err := os.WriteFile(path, []byte(content), 0o644); err != nil {
+			t.Fatalf("write %s: %v", rel, err)
+		}
+	}
+
+	mustWrite("data/agents.yml", `agents:
+  - owner: org
+    repo: small
+    tags: [terminal, community]
+  - owner: org
+    repo: big
+    tags: [web, vendor]
+    notes: a note
+`)
+
+	meta := metadataFile{
+		FetchedAt: fetchedAt,
+		Repos: map[string]repoMeta{
+			"org/big": {
+				NameWithOwner: "org/big", URL: "https://github.com/org/big",
+				Stars: 5000, Language: "Go", PushedAt: fetchedAt,
+				Description: "big one",
+			},
+			"org/small": {
+				NameWithOwner: "org/small", URL: "https://github.com/org/small",
+				Stars: 2000, Language: "Rust", PushedAt: fetchedAt,
+				Description: "small one", IsArchived: true,
+			},
+		},
+	}
+	raw, err := json.Marshal(meta)
+	if err != nil {
+		t.Fatalf("marshal metadata: %v", err)
+	}
+	mustWrite("data/metadata.json", string(raw))
+
+	// Baseline 7 days before fetchedAt (2026-09-09) so deltas resolve.
+	mustWrite("data/history.jsonl", `{"date":"2026-09-09","stars":{"org/big":4900,"org/small":1990}}
+{"date":"2026-09-16","stars":{"org/big":5000,"org/small":2000}}
+`)
+
+	mustWrite("site/index.html", "<html>dashboard</html>")
+	mustWrite("site/_headers", "/data.json\n  Cache-Control: no-cache\n")
+
+	return dir
+}
+
+func readBuiltSiteData(t *testing.T, dir string) siteData {
+	t.Helper()
+	raw, err := os.ReadFile(filepath.Join(dir, "dist", "data.json"))
+	if err != nil {
+		t.Fatalf("read dist/data.json: %v", err)
+	}
+	var got siteData
+	if err := json.Unmarshal(raw, &got); err != nil {
+		t.Fatalf("unmarshal dist/data.json: %v", err)
+	}
+	return got
+}
+
+func runBuildIn(t *testing.T, dir string) error {
+	t.Helper()
+	return runBuild(
+		filepath.Join(dir, "data/agents.yml"),
+		filepath.Join(dir, "data/metadata.json"),
+		filepath.Join(dir, "data/history.jsonl"),
+		filepath.Join(dir, "site"),
+		filepath.Join(dir, "dist"),
+	)
+}
+
+func TestRunBuild_JoinsCurationWithMetadataAndCopiesSite(t *testing.T) {
+	fetchedAt := time.Date(2026, 9, 16, 22, 10, 0, 0, time.UTC)
+	dir := writeBuildFixture(t, fetchedAt)
+
+	if err := runBuildIn(t, dir); err != nil {
+		t.Fatalf("runBuild: %v", err)
+	}
+
+	got := readBuiltSiteData(t, dir)
+
+	// updatedAt must label the fetch, not the build.
+	if got.UpdatedAt != "2026-09-16 22:10 UTC" {
+		t.Errorf("UpdatedAt: expected the metadata fetch time, got %q", got.UpdatedAt)
+	}
+
+	if len(got.Rows) != 2 {
+		t.Fatalf("expected 2 rows, got %d", len(got.Rows))
+	}
+	// Ranked by stars descending, independent of agents.yml order.
+	if got.Rows[0].Key != "org/big" || got.Rows[1].Key != "org/small" {
+		t.Errorf("expected big ranked above small, got %q then %q", got.Rows[0].Key, got.Rows[1].Key)
+	}
+	// Curation comes from agents.yml...
+	if got.Rows[0].Notes != "a note" {
+		t.Errorf("expected notes from agents.yml, got %q", got.Rows[0].Notes)
+	}
+	// ...figures from metadata.json.
+	if got.Rows[0].Stars != 5000 || got.Rows[0].Language != "Go" || got.Rows[0].Description != "big one" {
+		t.Errorf("row0 metadata not applied: %+v", got.Rows[0])
+	}
+	if !got.Rows[1].Archived {
+		t.Error("expected archived flag to survive the round trip")
+	}
+	if !got.Rows[0].HasDelta || got.Rows[0].Delta7d != 100 {
+		t.Errorf("row0 delta7d: expected +100 from the 2026-09-09 baseline, got hasDelta=%v delta=%d", got.Rows[0].HasDelta, got.Rows[0].Delta7d)
+	}
+	if len(got.Facets) == 0 {
+		t.Error("expected tag facets in the payload")
+	}
+
+	// Every static file in site/ must land in dist/ alongside data.json.
+	for _, name := range []string{"index.html", "_headers"} {
+		if _, err := os.Stat(filepath.Join(dir, "dist", name)); err != nil {
+			t.Errorf("expected dist/%s: %v", name, err)
+		}
+	}
+}
+
+// The build must not depend on the clock: it runs whenever Cloudflare happens
+// to redeploy, which may be long after the data was fetched.
+func TestRunBuild_DeltasAnchoredToFetchTimeNotWallClock(t *testing.T) {
+	fetchedAt := time.Date(2026, 9, 16, 22, 10, 0, 0, time.UTC)
+	dir := writeBuildFixture(t, fetchedAt)
+
+	orig := timeNow
+	// Simulate a redeploy 40 days later; a wall-clock anchor would slide the
+	// 7-day window past its slack and drop the delta entirely.
+	timeNow = func() time.Time { return fetchedAt.AddDate(0, 0, 40) }
+	defer func() { timeNow = orig }()
+
+	if err := runBuildIn(t, dir); err != nil {
+		t.Fatalf("runBuild: %v", err)
+	}
+
+	got := readBuiltSiteData(t, dir)
+	if !got.Rows[0].HasDelta || got.Rows[0].Delta7d != 100 {
+		t.Errorf("delta should be unchanged by a late rebuild, got hasDelta=%v delta=%d", got.Rows[0].HasDelta, got.Rows[0].Delta7d)
+	}
+	if got.UpdatedAt != "2026-09-16 22:10 UTC" {
+		t.Errorf("UpdatedAt should still report the fetch time, got %q", got.UpdatedAt)
+	}
+}
+
+// A newly contributed entry has no metadata until the next update run. The
+// build must degrade to omitting it, not fail — otherwise every deploy is
+// blocked between the PR merge and the nightly refresh.
+func TestRunBuild_SkipsEntriesWithoutMetadataYet(t *testing.T) {
+	fetchedAt := time.Date(2026, 9, 16, 22, 10, 0, 0, time.UTC)
+	dir := writeBuildFixture(t, fetchedAt)
+
+	agentsPath := filepath.Join(dir, "data/agents.yml")
+	existing, err := os.ReadFile(agentsPath)
+	if err != nil {
+		t.Fatalf("read agents.yml: %v", err)
+	}
+	added := string(existing) + "  - owner: org\n    repo: brandnew\n    tags: [terminal]\n"
+	if err := os.WriteFile(agentsPath, []byte(added), 0o644); err != nil {
+		t.Fatalf("write agents.yml: %v", err)
+	}
+
+	if err := runBuildIn(t, dir); err != nil {
+		t.Fatalf("runBuild should tolerate a missing metadata entry, got: %v", err)
+	}
+
+	got := readBuiltSiteData(t, dir)
+	if len(got.Rows) != 2 {
+		t.Fatalf("expected the un-fetched entry to be omitted, got %d rows", len(got.Rows))
+	}
+	for _, r := range got.Rows {
+		if r.Key == "org/brandnew" {
+			t.Error("org/brandnew has no metadata and must not be published")
+		}
+	}
+}
+
+// dist/ is build output: a file removed from site/ must not survive there.
+func TestRunBuild_WipesStaleDistFiles(t *testing.T) {
+	fetchedAt := time.Date(2026, 9, 16, 22, 10, 0, 0, time.UTC)
+	dir := writeBuildFixture(t, fetchedAt)
+
+	stale := filepath.Join(dir, "dist", "old.html")
+	if err := os.MkdirAll(filepath.Dir(stale), 0o755); err != nil {
+		t.Fatalf("mkdir dist: %v", err)
+	}
+	if err := os.WriteFile(stale, []byte("stale"), 0o644); err != nil {
+		t.Fatalf("write stale file: %v", err)
+	}
+
+	if err := runBuildIn(t, dir); err != nil {
+		t.Fatalf("runBuild: %v", err)
+	}
+
+	if _, err := os.Stat(stale); !os.IsNotExist(err) {
+		t.Errorf("expected stale dist file to be removed, stat err = %v", err)
+	}
+}
+
+func TestRunBuild_MissingMetadataIsActionable(t *testing.T) {
+	dir := t.TempDir()
+	if err := os.MkdirAll(filepath.Join(dir, "data"), 0o755); err != nil {
+		t.Fatalf("mkdir: %v", err)
+	}
+	if err := os.WriteFile(filepath.Join(dir, "data/agents.yml"), []byte("agents:\n  - owner: org\n    repo: big\n"), 0o644); err != nil {
+		t.Fatalf("write agents.yml: %v", err)
+	}
+
+	err := runBuildIn(t, dir)
+	if err == nil {
+		t.Fatal("expected an error when metadata.json is absent")
+	}
+	if !strings.Contains(err.Error(), "metadata.json") || !strings.Contains(err.Error(), "GITHUB_TOKEN") {
+		t.Errorf("error should name the missing file and how to produce it, got: %v", err)
+	}
+}

data/agents.yml

@@ -17,10 +17,7 @@ agents:
     tags: [terminal, byo-model, local-models, interactive, community]
   - owner: cline
     repo: cline
-    tags: [editor-plugin, terminal, byo-model, local-models, interactive, mcp, community]
+    tags: [editor-plugin, terminal, desktop, byo-model, local-models, interactive, headless, mcp, community]
-  - owner: continuedev
-    repo: continue
-    tags: [editor-plugin, terminal, byo-model, local-models, interactive, mcp, community]
   - owner: anomalyco
     repo: opencode
     tags: [terminal, desktop, byo-model, local-models, interactive, headless, mcp, community]
@@ -30,44 +27,32 @@ agents:
   - owner: google-gemini
     repo: gemini-cli
     tags: [terminal, single-vendor, interactive, headless, mcp, vendor]
-    notes: Google announced a closed-source successor; OSS repo may be retired
   - owner: openai
     repo: codex
-    tags: [terminal, editor-plugin, single-vendor, interactive, headless, mcp, vendor]
+    tags: [terminal, editor-plugin, desktop, single-vendor, interactive, headless, mcp, vendor]
   - owner: aaif-goose
     repo: goose
     tags: [terminal, desktop, byo-model, local-models, interactive, mcp, acp, community]
     notes: Moved to the Agentic AI Foundation (Linux Foundation)
   - owner: AntonOsika
     repo: gpt-engineer
-    tags: [terminal, byo-model, autonomous, community]
+    tags: [terminal, byo-model, local-models, autonomous, community]
     notes: Archived upstream (read-only); kept for historical significance
   - owner: SWE-agent
     repo: SWE-agent
     tags: [terminal, byo-model, autonomous, research, community]
-  - owner: RooCodeInc
-    repo: Roo-Code
-    tags: [editor-plugin, byo-model, interactive, mcp, community]
-    notes: Archived upstream (read-only)
   - owner: TabbyML
     repo: tabby
     tags: [self-hosted, editor-plugin, local-models, interactive, community]
   - owner: zed-industries
     repo: zed
     tags: [ide, byo-model, local-models, interactive, acp, community]
-  - owner: voideditor
-    repo: void
-    tags: [ide, byo-model, local-models, interactive, community]
-    notes: Archived upstream (read-only)
   - owner: anthropics
     repo: claude-code
-    tags: [terminal, editor-plugin, single-vendor, interactive, headless, mcp, vendor]
+    tags: [terminal, editor-plugin, desktop, single-vendor, interactive, headless, mcp, vendor]
   - owner: avante-corp
     repo: avante.nvim
     tags: [editor-plugin, byo-model, local-models, interactive, mcp, acp, community]
-  - owner: kortix-ai
-    repo: suna
-    tags: [web, self-hosted, byo-model, autonomous, mcp, community]
   - owner: earendil-works
     repo: pi
     tags: [terminal, byo-model, interactive, community]
@@ -83,10 +68,10 @@ agents:
     tags: [terminal, editor-plugin, byo-model, autonomous, community]
   - owner: QwenLM
     repo: qwen-code
-    tags: [terminal, editor-plugin, desktop, byo-model, local-models, interactive, headless, mcp, acp, vendor]
+    tags: [terminal, editor-plugin, desktop, web, byo-model, local-models, interactive, headless, mcp, acp, vendor]
   - owner: Kilo-Org
     repo: kilocode
-    tags: [editor-plugin, terminal, byo-model, interactive, mcp, community]
+    tags: [editor-plugin, terminal, byo-model, interactive, autonomous, headless, mcp, community]
   - owner: onlook-dev
     repo: onlook
     tags: [desktop, web, byo-model, app-builder, mcp, community]
@@ -98,7 +83,7 @@ agents:
     tags: [terminal, single-vendor, interactive, headless, mcp, vendor]
   - owner: deepseek-ai
     repo: deepseek-harness
-    tags: [terminal, single-vendor, interactive, vendor]
+    tags: [terminal, web, single-vendor, interactive, vendor]
     notes: Developer preview; compatibility-breaking changes expected
   - owner: openinterpreter
     repo: openinterpreter
@@ -108,10 +93,10 @@ agents:
     tags: [terminal, editor-plugin, desktop, byo-model, local-models, interactive, mcp, community]
   - owner: esengine
     repo: DeepSeek-Reasonix
-    tags: [terminal, desktop, web, editor-plugin, byo-model, interactive, autonomous, mcp, acp, community]
+    tags: [terminal, desktop, web, editor-plugin, single-vendor, interactive, autonomous, mcp, acp, community]
   - owner: Gitlawb
     repo: openclaude
-    tags: [terminal, byo-model, local-models, interactive, headless, mcp, community]
+    tags: [terminal, editor-plugin, byo-model, local-models, interactive, headless, mcp, community]
   - owner: can1357
     repo: oh-my-pi
     tags: [terminal, editor-plugin, byo-model, local-models, interactive, mcp, acp, headless, community]
@@ -120,7 +105,7 @@ agents:
     tags: [terminal, editor-plugin, single-vendor, interactive, headless, mcp, acp, vendor]
   - owner: alibaba
     repo: open-code-review
-    tags: [terminal, byo-model, review, mcp, vendor]
+    tags: [terminal, byo-model, review, headless, mcp, vendor]
   - owner: PrimeIntellect-ai
     repo: prime-agent
     tags: [terminal, byo-model, autonomous, headless, vendor]
@@ -139,15 +124,11 @@ agents:
   - owner: Kuberwastaken
     repo: claurst
     tags: [terminal, byo-model, interactive, headless, acp, community]
-  - owner: Orkas-AI
-    repo: Orkas
-    tags: [desktop, byo-model, interactive, community]
-    notes: "Open-source, local-first desktop AI workforce coordinated by a Commander through one chat; supports local coding-agent workflows."
   - owner: stablyai
     repo: orca
     tags: [desktop, terminal, byo-model, orchestration, headless, community]
     notes: "ADE: runs Codex, Claude Code, OpenCode and Pi side-by-side, each in its own git worktree, with a mobile companion app."
   - owner: getpaseo
     repo: paseo
-    tags: [desktop, web, terminal, self-hosted, byo-model, orchestration, headless, community]
+    tags: [desktop, web, terminal, editor-plugin, self-hosted, byo-model, orchestration, headless, mcp, community]
     notes: "ADE: self-hosted daemon that runs Claude Code, Codex, Copilot, OpenCode and Pi on your own machines, driven from desktop, mobile, web or CLI."

docs/CONTRIBUTING.md

@@ -41,8 +41,11 @@ The dividing line is **tools you use vs. building blocks you import**. Out of sc
 libraries and SDKs, agent frameworks meant to be built on, model weights, prompt or
 skill collections, and dashboards that only observe a tool without being one.
 
-A tool also has to be *about* software development. General-purpose assistants and
+A tool also has to be *about* software development. General-purpose assistants,
-chat UIs do not qualify just because a developer could use them.
+chat UIs, and multi-agent "digital workforce" apps — the ones whose specialists write
+reports, decks, and marketing copy — do not qualify just because a developer could use
+them, or because one of their agents happens to touch code. Judge the tool by what it is
+built to do, not by the widest thing it can be pointed at.
 
 ### The star floor is hard
 
@@ -63,7 +66,7 @@ describe a tool that ships as a CLI, an editor plugin and a desktop app at the
 same time — which most of them now do. An entry carries several tags across
 five facets. `data/agents.yml` is the source of truth; the vocabulary itself
 lives in `tagVocabulary` (`validate.go`) and reaches the dashboard through
-`site/data.json`, so it is defined exactly once.
+the generated `dist/data.json`, so it is defined exactly once.
 
 **Surface** — where you run it. At least one required.
 
@@ -121,6 +124,9 @@ say nothing about choosing the tool. GitHub topics are a drafting aid only —
 **Deprecation:** To remove an agent, delete its entry from `data/agents.yml`. The next run drops it from the README; its history stays in `data/history.jsonl`, so re-adding the entry later restores its star chart.
 
 **Staleness:** An entry with no push in **6 months** is dropped. Past **3 months** the updater prints a `::warning::` naming the repo and its days idle, so the daily run surfaces candidates without anyone auditing the list by hand. Removal stays a human decision: a repo can go quiet between releases, and a historically significant one (`gpt-engineer`) is kept with a `notes` marker instead.
+A repo the maintainers have **archived or declared deprecated** is different: it will never
+be pushed again, so it is removed without waiting out the 6 months unless it earns the same
+historical-significance exception.
 
 ## PR Review
 
@@ -128,4 +134,5 @@ say nothing about choosing the tool. GitHub topics are a drafting aid only —
 - The daily GitHub Actions workflow (runs at 00:00 UTC) picks up merged PRs automatically
 - No manual review required; the updater regenerates the README after your PR merges
 
-For local testing before opening a PR, see [LOCAL_DEV.md](./LOCAL_DEV.md).
+For local testing before opening a PR, see [LOCAL_DEV.md](./LOCAL_DEV.md). A tag
+or note change needs only `go run . -build` — no GitHub token.

docs/DEPLOY.md

@@ -0,0 +1,99 @@
+# Deploying to Cloudflare Pages
+
+The site is published by Cloudflare Pages' Git integration. Cloudflare builds
+from the repository on every push to `main` and never holds a GitHub token.
+
+## How the split works
+
+Data refresh and site build are two separate steps, in two separate places:
+
+| Step | Command | Runs where | Needs a token? | Writes |
+| --- | --- | --- | --- | --- |
+| Update | `go run .` | GitHub Actions (`update.yml`, nightly + manual) | Yes — `GITHUB_TOKEN` | `README.md`, `data/history.jsonl`, `data/metadata.json` (all committed) |
+| Build | `go run . -build` | Cloudflare Pages | No | `dist/` (never committed) |
+
+`data/metadata.json` is the handoff. The update step records every GitHub-sourced
+field there — stars, language, description, push date, archived flag — so the
+build step can render the dashboard from committed files alone.
+
+That has two consequences worth knowing:
+
+- Cloudflare's build environment never sees a GitHub token, because it has no
+  reason to call the GitHub API.
+- A pure curation change (retagging an entry, editing a note in
+  `data/agents.yml`) republishes as soon as you push, reusing the last fetched
+  star figures. You do not wait for the nightly run.
+
+The nightly Actions run commits refreshed data to `main`; that push fires
+Cloudflare's build webhook, which redeploys with the new numbers.
+
+## One-time setup
+
+### 1. Bootstrap `data/metadata.json`
+
+The build fails without it, so generate it before connecting Cloudflare. Either
+trigger the **Update rankings** workflow manually (Actions tab →
+*Update rankings* → *Run workflow*), or run the updater locally and commit:
+
+```bash
+export GITHUB_TOKEN=ghp_your_token_here
+go run .
+git add data/metadata.json data/history.jsonl README.md
+git commit -m "chore: bootstrap metadata snapshot"
+git push
+```
+
+### 2. Create the Pages project
+
+Cloudflare dashboard → **Workers & Pages** → **Create** → **Pages** →
+**Connect to Git** → pick this repository, then set:
+
+| Setting | Value |
+| --- | --- |
+| Production branch | `main` |
+| Framework preset | None |
+| Build command | `go run . -build` |
+| Build output directory | `dist` |
+| Root directory | *(leave blank)* |
+
+### 3. Set the build environment variable
+
+Under **Settings → Environment variables → Production** (and Preview, if you
+want PR previews) add:
+
+| Variable | Value |
+| --- | --- |
+| `GO_VERSION` | `1.23` |
+
+Do **not** add `GITHUB_TOKEN` — the build does not use one, and adding it would
+hand a credential to an environment that has no need for it.
+
+Cloudflare's build image ships Go and honours `GO_VERSION`. Even on an older
+image, Go's toolchain directive in `go.mod` downloads the matching toolchain
+automatically.
+
+### 4. Deploy
+
+Save and deploy. Subsequent pushes to `main` — yours and the nightly bot's —
+redeploy automatically.
+
+## Caching
+
+`site/_headers` marks `data.json` as `must-revalidate`, so the edge cannot serve
+yesterday's ranking after a refresh. `index.html` and everything else in `site/`
+are copied into `dist/` as-is and use Cloudflare's defaults.
+
+## Troubleshooting
+
+**Build fails with "data/metadata.json not found"** — the bootstrap in step 1
+has not been committed yet.
+
+**A newly added tool is missing from the site** — expected between merging the
+`agents.yml` entry and the next update run. The build logs a warning and omits
+entries it has no metadata for, rather than failing the deploy. Trigger the
+*Update rankings* workflow manually to fetch it immediately.
+
+**Stars look stale** — check the Actions tab: publishing is healthy, but the
+update workflow has not committed recently. The dashboard's "updated" timestamp
+reports when the data was fetched, not when the site was built, so a redeploy
+never makes stale figures look fresh.

docs/LOCAL_DEV.md

@@ -9,25 +9,56 @@
 
 1. Go to [https://github.com/settings/tokens](https://github.com/settings/tokens)
 2. Click "Generate new token" → "Generate new token (classic)"
-3. Give it a name (e.g., "ai-dev-tools")
+3. Give it a name (e.g., "awesome-ai-dev-tools")
 4. Select scope: **`public_repo`** (needed to read public repo metadata)
 5. Click "Generate token" and copy the value
 
 The updater reads public repos only; `public_repo` scope is sufficient and safe.
 
-## Running Locally
+## The two steps
+
+The tool separates fetching data from rendering the site, so only the fetch
+needs a token. See [DEPLOY.md](./DEPLOY.md) for why.
+
+### Update (needs a token, hits the network)
 
 ```bash
 export GITHUB_TOKEN=ghp_your_token_here
 go run .
 ```
 
-The tool will:
+This will:
 - Read `data/agents.yml`
-- Fetch live metadata from GitHub GraphQL API
+- Fetch live metadata from the GitHub GraphQL API
 - Append star counts to `data/history.jsonl`
+- Write the fetched fields to `data/metadata.json`
 - Regenerate `README.md` from `templates/readme.tmpl`
 
+### Build (offline, no token)
+
+```bash
+go run . -build
+```
+
+This joins `data/agents.yml` (tags and notes) with `data/metadata.json` (stars
+and repo metadata) and renders `dist/` — a copy of `site/` plus the generated
+`dist/data.json`. This is exactly what Cloudflare Pages runs.
+
+Preview it with any static server:
+
+```bash
+go run . -build && python3 -m http.server -d dist 8080
+```
+
+If you only touched `site/index.html` or the tags in `data/agents.yml`, the
+build step alone is enough — no token required.
+
+### Validate (offline, no token)
+
+```bash
+go run . -check
+```
+
 ## Reverting Local Changes
 
 Before opening a PR, undo local modifications:
@@ -48,3 +79,4 @@ Always set `GITHUB_TOKEN` when testing locally to avoid hitting the unauthentica
 ## Next Steps
 
 To add agents, see [CONTRIBUTING.md](./CONTRIBUTING.md).
+To publish the site, see [DEPLOY.md](./DEPLOY.md).

github.go

@@ -156,16 +156,21 @@ func fetchStats(token string, agents []Agent) ([]Stat, error) {
 		})
 	}
 
-	// Sort by stars descending. Ties are ordered by CanonicalKey for determinism
+	sortStats(stats)
-	// regardless of map-iteration or agents.yml order.
+
+	return stats, nil
+}
+
+// sortStats orders the ranking by stars descending. Ties break on CanonicalKey
+// for determinism regardless of map-iteration or agents.yml order. Shared with
+// the build step, which re-ranks from committed metadata.
+func sortStats(stats []Stat) {
 	sort.Slice(stats, func(i, j int) bool {
 		if stats[i].Stars != stats[j].Stars {
 			return stats[i].Stars > stats[j].Stars
 		}
 		return stats[i].CanonicalKey < stats[j].CanonicalKey
 	})
-
-	return stats, nil
 }
 
 // fetchChunk sends one GraphQL request for a slice of agents. aliasOffset
@@ -251,7 +256,7 @@ func doWithRetry(token string, body []byte) ([]byte, int, error) {
 		}
 		req.Header.Set("Authorization", "Bearer "+token)
 		req.Header.Set("Content-Type", "application/json")
-		req.Header.Set("User-Agent", "ai-dev-tools-updater")
+		req.Header.Set("User-Agent", "awesome-ai-dev-tools-updater")
 
 		resp, err := httpClient.Do(req)
 		if err != nil {

go.mod

@@ -1,4 +1,4 @@
-module github.com/tiennm99/ai-dev-tools
+module github.com/tiennm99/awesome-ai-dev-tools
 
 go 1.23
 

history.go

@@ -195,8 +195,16 @@ func computeDeltas(history []Snapshot, current Snapshot) map[string]int {
 // the delta would be misleadingly labeled "Δ<days>d", so we return no delta
 // for that repo instead.
 func computeDeltaOver(history []Snapshot, current Snapshot, days, slackDays int) map[string]int {
+	return computeDeltaAt(history, current, timeNow(), days, slackDays)
+}
+
+// computeDeltaAt is computeDeltaOver with an explicit anchor instead of the
+// wall clock. The build step passes the time the data was fetched, so
+// rebuilding an unchanged dataset later reproduces identical deltas rather
+// than sliding the window forward off its baseline.
+func computeDeltaAt(history []Snapshot, current Snapshot, anchor time.Time, days, slackDays int) map[string]int {
 	deltas := map[string]int{}
-	now := timeNow().UTC()
+	now := anchor.UTC()
 	cutoff := now.AddDate(0, 0, -days).Format("2006-01-02")
 	lowerBound := now.AddDate(0, 0, -(days + slackDays)).Format("2006-01-02")
 

main.go

@@ -11,28 +11,52 @@ import (
 const (
 	agentsPath   = "data/agents.yml"
 	historyPath  = "data/history.jsonl"
+	metadataPath = "data/metadata.json"
 	readmeTmpl   = "templates/readme.tmpl"
 	readmePath   = "README.md"
-	siteDataPath = "site/data.json"
+	siteDir      = "site"
+	distDir      = "dist"
 )
 
+// The tool has three modes, deliberately separated so the only step that needs
+// network access and a GITHUB_TOKEN is the one that runs in CI:
+//
+//	(default)  update — fetch GitHub, refresh history/metadata/README
+//	-build            — render dist/ from committed data, offline
+//	-check            — validate data/agents.yml, offline
+//
+// The split is what lets an untrusted build environment (Cloudflare Pages)
+// publish the site without ever holding a token.
 func main() {
 	check := flag.Bool("check", false, "validate data/agents.yml offline (no network, no token) and exit")
+	build := flag.Bool("build", false, "render the static site into dist/ from committed data (no network, no token) and exit")
 	flag.Parse()
 
-	if *check {
+	if *check && *build {
+		log.Print("-check and -build are mutually exclusive")
+		os.Exit(1)
+	}
+
+	switch {
+	case *check:
 		if err := runCheck(agentsPath); err != nil {
 			log.Printf("check failed: %v", err)
 			os.Exit(1)
 		}
-		return
+	case *build:
+		if err := runBuild(agentsPath, metadataPath, historyPath, siteDir, distDir); err != nil {
+			log.Printf("build failed: %v", err)
+			os.Exit(1)
 		}
-
+	default:
 		if err := run(); err != nil {
 			log.Fatalf("update failed: %v", err)
 		}
+	}
 }
 
+// run is the update step: it is the only mode that talks to GitHub. It writes
+// data/metadata.json, which the build step later consumes offline.
 func run() error {
 	agents, err := loadAgents(agentsPath)
 	if err != nil {
@@ -57,7 +81,7 @@ func run() error {
 		return fmt.Errorf("no agents in %s meet the %d-star minimum", agentsPath, minStars)
 	}
 
-	snapshots, deltas7, deltas30, err := appendHistory(historyPath, stats)
+	_, deltas7, _, err := appendHistory(historyPath, stats)
 	if err != nil {
 		return err
 	}
@@ -66,7 +90,7 @@ func run() error {
 		return err
 	}
 
-	if err := writeSiteData(siteDataPath, stats, deltas7, deltas30, snapshots); err != nil {
+	if err := writeMetadata(metadataPath, stats); err != nil {
 		return err
 	}
 

metadata.go

@@ -0,0 +1,112 @@
+package main
+
+import (
+	"encoding/json"
+	"fmt"
+	"io"
+	"os"
+	"time"
+)
+
+// repoMeta is the GitHub-sourced half of a ranking row: every field the
+// dashboard needs that cannot be derived from data/agents.yml. It is written
+// by the update step (which has a token) and committed, so the build step can
+// render the site offline.
+type repoMeta struct {
+	NameWithOwner string    `json:"nameWithOwner"`
+	URL           string    `json:"url"`
+	Stars         int       `json:"stars"`
+	Language      string    `json:"language"`
+	PushedAt      time.Time `json:"pushedAt"`
+	Description   string    `json:"description"`
+	IsArchived    bool      `json:"archived"`
+}
+
+// metadataFile is the on-disk shape of data/metadata.json. Repos is keyed by
+// canonical owner/repo (the same key history.jsonl uses), so entries survive
+// renames and Go's sorted map marshalling keeps commit diffs minimal.
+type metadataFile struct {
+	FetchedAt time.Time           `json:"fetchedAt"`
+	Repos     map[string]repoMeta `json:"repos"`
+}
+
+// writeMetadata persists the fetched GitHub metadata. Unlike the old
+// site/data.json this file IS committed: it is the handoff between the update
+// step and the build step, and the only reason the build can run without a
+// GITHUB_TOKEN.
+func writeMetadata(path string, stats []Stat) error {
+	repos := make(map[string]repoMeta, len(stats))
+	for _, s := range stats {
+		repos[s.CanonicalKey] = repoMeta{
+			NameWithOwner: s.NameWithOwner,
+			URL:           s.URL,
+			Stars:         s.Stars,
+			Language:      s.Language,
+			PushedAt:      s.PushedAt,
+			Description:   s.Description,
+			IsArchived:    s.IsArchived,
+		}
+	}
+
+	return atomicWriteFile(path, func(w io.Writer) error {
+		enc := json.NewEncoder(w)
+		// Indented so the committed file reviews as a readable diff.
+		enc.SetIndent("", "  ")
+		return enc.Encode(metadataFile{
+			FetchedAt: timeNow().UTC(),
+			Repos:     repos,
+		})
+	})
+}
+
+func readMetadata(path string) (metadataFile, error) {
+	var m metadataFile
+	data, err := os.ReadFile(path)
+	if err != nil {
+		if os.IsNotExist(err) {
+			return m, fmt.Errorf("%s not found — run the updater (`go run .` with GITHUB_TOKEN set) to generate it", path)
+		}
+		return m, err
+	}
+	if err := json.Unmarshal(data, &m); err != nil {
+		return m, fmt.Errorf("parse %s: %w", path, err)
+	}
+	if len(m.Repos) == 0 {
+		return m, fmt.Errorf("%s contains no repos", path)
+	}
+	return m, nil
+}
+
+// statsFromMetadata rejoins the two halves of a row: curation (order-independent
+// tags and notes) from agents.yml, live figures from metadata.json.
+//
+// A repo present in agents.yml but absent from metadata.json is skipped with a
+// warning rather than failing the build. That is the normal state between
+// "contributor adds an entry" and "the next nightly update fetches it", and a
+// hard failure there would block every Cloudflare deploy in the meantime.
+func statsFromMetadata(agents []Agent, meta metadataFile) []Stat {
+	stats := make([]Stat, 0, len(agents))
+	for _, a := range agents {
+		key := a.Owner + "/" + a.Repo
+		m, ok := meta.Repos[key]
+		if !ok {
+			fmt.Fprintf(os.Stderr, "warning: %s has no entry in metadata yet — omitted from this build; the next updater run will add it\n", key)
+			continue
+		}
+		stats = append(stats, Stat{
+			CanonicalKey:  key,
+			Owner:         a.Owner,
+			Repo:          a.Repo,
+			Tags:          a.Tags,
+			Notes:         a.Notes,
+			Description:   m.Description,
+			Stars:         m.Stars,
+			Language:      m.Language,
+			PushedAt:      m.PushedAt,
+			URL:           m.URL,
+			NameWithOwner: m.NameWithOwner,
+			IsArchived:    m.IsArchived,
+		})
+	}
+	return stats
+}

plans/reports/research-260911-1328-agents-refresh.md

@@ -1,6 +1,6 @@
-# Research Report: awesome-coding-agents list refresh (additions + stale cleanup)
+# Research Report: awesome-ai-dev-tools list refresh (additions + stale cleanup)
 
-**Conducted:** 2026-09-11 13:28 (+07) · **Repo:** `tiennm99/awesome-coding-agents` · **Tracked now:** 29
+**Conducted:** 2026-09-11 13:28 (+07) · **Repo:** `tiennm99/awesome-ai-dev-tools` · **Tracked now:** 29
 
 ## Executive Summary
 

site.go

@@ -5,7 +5,7 @@ import (
 	"io"
 )
 
-// siteRow is one ranked repo in site/data.json, consumed by site/index.html.
+// siteRow is one ranked repo in the generated dist/data.json, consumed by site/index.html.
 type siteRow struct {
 	Key           string   `json:"key"` // canonical owner/repo, matches history keys
 	NameWithOwner string   `json:"nameWithOwner"`
@@ -32,10 +32,13 @@ type siteData struct {
 	Facets []tagFacet `json:"facets"`
 }
 
-// writeSiteData emits the JSON payload for the GitHub Pages dashboard.
+// writeSiteData emits the JSON payload the dashboard fetches at runtime.
-// The file is generated fresh on every updater run and is not committed;
+// It is build output (dist/data.json), never committed.
-// the Pages deploy step in the workflow picks it up from the working tree.
+//
-func writeSiteData(path string, stats []Stat, deltas7, deltas30 map[string]int, history []Snapshot) error {
+// updatedAt is passed in rather than read from the clock because it labels
+// data freshness, not build time: a redeploy of unchanged data must not claim
+// the figures are newer than the fetch that produced them.
+func writeSiteData(path string, updatedAt string, stats []Stat, deltas7, deltas30 map[string]int, history []Snapshot) error {
 	rows := make([]siteRow, len(stats))
 	for i, s := range stats {
 		delta7, has7 := deltas7[s.CanonicalKey]
@@ -60,7 +63,7 @@ func writeSiteData(path string, stats []Stat, deltas7, deltas30 map[string]int,
 
 	return atomicWriteFile(path, func(w io.Writer) error {
 		return json.NewEncoder(w).Encode(siteData{
-			UpdatedAt: timeNow().UTC().Format("2006-01-02 15:04 UTC"),
+			UpdatedAt: updatedAt,
 			Rows:      rows,
 			History:   history,
 			Facets:    tagVocabulary,

site/_headers

@@ -0,0 +1,5 @@
+# data.json is regenerated on every deploy and read by the dashboard at load.
+# Cloudflare's edge would otherwise keep serving yesterday's ranking after a
+# nightly refresh; the page's own `cache: 'no-cache'` only covers the browser.
+/data.json
+  Cache-Control: public, max-age=0, must-revalidate

site/index.html

@@ -3,11 +3,11 @@
 <head>
 <meta charset="utf-8">
 <meta name="viewport" content="width=device-width, initial-scale=1">
-<title>AI Dev Tools — Rankings &amp; History</title>
+<title>Awesome AI Dev Tools — Rankings &amp; History</title>
 <meta name="description" content="Daily-refreshed ranking of open-source AI coding agents by GitHub stars, with 7-day deltas and star-history charts.">
-<meta property="og:title" content="AI Dev Tools — Rankings &amp; History">
+<meta property="og:title" content="Awesome AI Dev Tools — Rankings &amp; History">
 <meta property="og:description" content="Daily-refreshed ranking of open-source AI coding agents by GitHub stars, with 7-day deltas and star-history charts.">
-<meta property="og:url" content="https://tiennm99.github.io/ai-dev-tools/">
+<meta property="og:url" content="https://awesome-ai-dev-tools.pages.dev/">
 <meta property="og:type" content="website">
 <meta name="theme-color" content="#f8fafc" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#0f172a" media="(prefers-color-scheme: dark)">
@@ -211,11 +211,11 @@
     <div>
       <h1>
         <svg width="22" height="22" viewBox="0 0 24 24" fill="currentColor" aria-hidden="true"><path d="M12 2l2.9 6.26L21.5 9.1l-4.9 4.5 1.3 6.7L12 17l-5.9 3.3 1.3-6.7-4.9-4.5 6.6-.84L12 2z"/></svg>
-        AI Dev Tools
+        Awesome AI Dev Tools
       </h1>
       <div class="sub">Curated ranking of open-source AI coding agents by GitHub stars, refreshed daily. <span id="meta"></span></div>
     </div>
-    <a class="ghlink" href="https://github.com/tiennm99/ai-dev-tools">
+    <a class="ghlink" href="https://github.com/tiennm99/awesome-ai-dev-tools">
       <svg width="18" height="18" viewBox="0 0 16 16" fill="currentColor" aria-hidden="true"><path d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27s1.36.09 2 .27c1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.01 8.01 0 0 0 16 8c0-4.42-3.58-8-8-8z"/></svg>
       GitHub
     </a>
@@ -225,7 +225,7 @@
   <noscript>
     <p>This dashboard needs JavaScript to render the ranking table and charts.
     You can view the same ranking in the
-    <a href="https://github.com/tiennm99/ai-dev-tools#readme">project README on GitHub</a>.</p>
+    <a href="https://github.com/tiennm99/awesome-ai-dev-tools#readme">project README on GitHub</a>.</p>
   </noscript>
 
   <div class="stats" id="stats" hidden>
@@ -272,8 +272,8 @@
   </section>
 
   <footer>
-    <span>Data refreshed daily by <a href="https://github.com/tiennm99/ai-dev-tools/actions">GitHub Actions</a>.</span>
+    <span>Data refreshed daily by <a href="https://github.com/tiennm99/awesome-ai-dev-tools/actions">GitHub Actions</a>.</span>
-    <span>Add an agent via <a href="https://github.com/tiennm99/ai-dev-tools#contributing">a pull request</a>.</span>
+    <span>Add an agent via <a href="https://github.com/tiennm99/awesome-ai-dev-tools#contributing">a pull request</a>.</span>
   </footer>
 
 <script src="https://cdn.jsdelivr.net/npm/chart.js@4.4.9/dist/chart.umd.min.js"></script>
@@ -309,7 +309,7 @@
   const isStale = r => r.pushedAt && new Date(r.pushedAt) < staleCutoff;
 
   // --- facet filter chips (multi-select) ---
-  // Vocabulary comes from site/data.json, which the Go updater fills from
+  // Vocabulary comes from the generated data.json, which the Go build step fills from
   // tagVocabulary — no second copy of the tag list lives in this file.
   const facets = Array.isArray(data.facets) ? data.facets : [];
   // Only offer a tag some row actually carries, so the rows never disagree

site_test.go

@@ -51,7 +51,7 @@ func TestWriteSiteData_JSONShapeAndDeltaFields(t *testing.T) {
 	tmpDir := t.TempDir()
 	tmpFile := tmpDir + "/data.json"
 
-	if err := writeSiteData(tmpFile, stats, deltas7, deltas30, history); err != nil {
+	if err := writeSiteData(tmpFile, "2026-08-09 12:00 UTC", stats, deltas7, deltas30, history); err != nil {
 		t.Fatalf("writeSiteData: %v", err)
 	}
 

templates/readme.tmpl

@@ -1,9 +1,9 @@
-# AI Dev Tools
+# Awesome AI Dev Tools
 
 > Curated ranking of AI-powered developer tools, sorted by GitHub stars.
 > Updated daily by GitHub Actions.
 
-📊 **[Interactive dashboard with star-history charts →](https://tiennm99.github.io/ai-dev-tools/)**
+📊 **[Interactive dashboard with star-history charts →](https://awesome-ai-dev-tools.pages.dev/)**
 
 **Last updated:** {{.UpdatedAt}} · **Tracked:** {{.Total}} repos
 {{- if .TopMover.HasMover }}
@@ -24,8 +24,8 @@
 2. A daily GitHub Actions workflow (`.github/workflows/update.yml`) runs the Go updater.
 3. The updater fetches live repo metadata via the GitHub GraphQL API in one batched query.
 4. Star counts are appended to `data/history.jsonl` for 7-day delta computation.
-5. This `README.md` is regenerated from `templates/readme.tmpl` and committed back to the repo.
+5. This `README.md` and `data/metadata.json` are regenerated and committed back to the repo.
-6. `site/data.json` is regenerated and the [dashboard](https://tiennm99.github.io/ai-dev-tools/) (`site/index.html`) is redeployed to GitHub Pages.
+6. That push triggers Cloudflare Pages, which runs `go run . -build` to render the [dashboard](https://awesome-ai-dev-tools.pages.dev/) from the committed data — no API token needed at build time. See [docs/DEPLOY.md](./docs/DEPLOY.md).
 
 **Δ7d:** Change in stars over the past 7 days; `—` means fewer than 7 days of history.
 
@@ -33,7 +33,7 @@
 
 **Inclusion criteria** — a repo belongs on this list when it is:
 
-1. **A developer tool built around AI** — something a developer uses to build software, where an LLM is central to what it does. That covers agents that write, edit, or review code; agent development environments that run and coordinate them; and AI-assisted editors, terminals, and review tools. Libraries, SDKs, model weights, and prompt or skill collections are out of scope — this list ranks tools you *use*, not building blocks you *import*.
+1. **A developer tool built around AI** — something a developer uses to build software, where an LLM is central to what it does. That covers agents that write, edit, or review code; agent development environments that run and coordinate them; and AI-assisted editors, terminals, and review tools. Libraries, SDKs, model weights, and prompt or skill collections are out of scope — this list ranks tools you *use*, not building blocks you *import*. General-purpose assistants and multi-agent "digital workforce" apps are out as well, even when one of their agents can write code — the tool itself has to be aimed at building software.
 2. **At least 1,000 GitHub stars** — a hard floor, enforced by the updater. An entry below it is dropped from the ranking automatically, so the threshold is never waived for an individual entry.
 3. **Open source and actively maintained** — no push in **6 months** means the entry is dropped. The daily run flags anything past **3 months** for review, so the ranking reflects tools people can actually use today.
 
@@ -48,7 +48,7 @@ agents:
     tags: [terminal, byo-model, interactive, community]
 ```
 
-**Tags** describe a tool across five facets — the [dashboard](https://tiennm99.github.io/ai-dev-tools/) filters on them, OR within a facet and AND across facets:
+**Tags** describe a tool across five facets — the [dashboard](https://awesome-ai-dev-tools.pages.dev/) filters on them, OR within a facet and AND across facets:
 
 | Facet | Tags |
 |-------|------|

validate.go

@@ -28,7 +28,7 @@ var repoPattern = regexp.MustCompile(`^[A-Za-z0-9._-]+$`)
 //
 // This slice is the single source of truth: the lookup map below, the
 // validation messages, and the dashboard's filter chips (shipped in
-// site/data.json) are all derived from it.
+// the generated dist/data.json) are all derived from it.
 var tagVocabulary = []tagFacet{
 	{ID: facetSurface, Label: "Surface", Tags: []string{"terminal", "editor-plugin", "ide", "desktop", "web", "self-hosted"}},
 	{ID: facetModel, Label: "Model access", Tags: []string{"byo-model", "single-vendor", "local-models"}},