Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

docs(growth): record four-repo homepage routing (#3509)

zhifu gao4a58b27293e5
diff --git a/docs/community_growth_20k.md b/docs/community_growth_20k.md
index d45ac0234..66f984329 100644
--- a/docs/community_growth_20k.md
+++ b/docs/community_growth_20k.md
@@ -64,6 +64,13 @@ As of 2026-08-17 00:48 UTC, the ecosystem has 36,584 combined GitHub stars, or 5
- The release audit exposed a stale patrol expectation on the two llama.cpp comparison articles: the pages correctly referenced `runtime-llamacpp-v0.2.0`, while the checker still required `v0.1.9`. FunASR PR [#3504](https://github.com/modelscope/FunASR/pull/3504) merged as `0d663c6d7fe64aecb117c0337abfab526c1733cb`, added a product-build contract test, and restored the live patrol to 20 page contracts, 98 navigation pages, and 3 static assets passing.
- Bilingual announcement [#3506](https://github.com/modelscope/FunASR/discussions/3506) now gives OpenMAIC users copy-paste CPU SenseVoice and NVIDIA Fun-ASR-Nano commands for `funasr==1.4.2`, the exact `ASR_FUNASR_BASE_URL=http://localhost:8000/v1` contract, and the adapter's WAV-only boundary. The published body was read back and matched against the backed-up source, and the merged OpenMAIC PR links to the official catalog and announcement in [the upstream follow-up](https://github.com/THU-MAIC/OpenMAIC/pull/1044#issuecomment-5310622587).
+### 2026-08-17 four-repository homepage routing
+
+- The refreshed 14-day GitHub traffic window measured FunASR at 36,728 views / 8,881 unique visitors, Fun-ASR at 4,580 / 1,499, SenseVoice at 8,315 / 2,996, and FunClip at 3,408 / 1,520. In the same repository order, `funasr.com` sent 2,597 / 724, 575 / 231, 181 / 87, and 96 / 43 referrer visits / unique visitors. The owned site is already a material FunASR discovery source but under-routes its traffic to the other three repositories.
+- FunASR PR [#3508](https://github.com/modelscope/FunASR/pull/3508) merged as `2cb615c42911972602cfaec739144d406b3282c5`. The bilingual homepage now presents all four repositories in stable order and labels their workloads as Production speech toolkit, High-accuracy ASR, Languages and emotion, and AI video editing; each entry uses its attributable `/go/*` route.
+- Local evidence passed 109 Python tests and 22 Playwright tests, including desktop and mobile order, visibility, row-layout, overflow, console, and network assertions. Two clean builds each generated 27 product pages and validated 102 public pages with no diff; exact-head GitHub Actions passed both required checks.
+- Production release `20260817T012430Z` was built twice from the exact merge commit, atomically replaced `20260817T004144Z`, and retained that previous release for rollback. The 7,175,781-byte archive has SHA-256 `5c5b96595d4cb893c83aedeca4d5dfe5a4cf0a665a1b4d489808886a7aacc1a3`; preflight, staging, post-switch, Chinese and English public-page, all four redirect, nginx, cache-header, and security-header checks passed.
+
### 2026-08-16 OpenClaw realtime transcription SDK unblock
- OpenClaw PR [#118977](https://github.com/openclaw/openclaw/pull/118977) was synchronized with `openclaw/main@2d3612da6bd28f6a7956d82b0795c31fd18bab8a`, resolving its only merge conflict while preserving both upstream SDK surface-budget changes and the FunASR WebSocket subprotocol work.
diff --git a/tests/test_growth_plan_doc.py b/tests/test_growth_plan_doc.py
index 4a177ae79..ac259ba3b 100644
--- a/tests/test_growth_plan_doc.py
+++ b/tests/test_growth_plan_doc.py
@@ -401,3 +401,30 @@ def test_growth_plan_records_metrics_tracking_completion():
assert "[ ] Track stars, PyPI downloads, issue volume, and docs traffic." not in text
assert "docs traffic remains unavailable" not in text
+
+
+def test_growth_plan_records_four_repo_homepage_routing_release():
+ text = PLAN.read_text()
+
+ required_markers = [
+ "### 2026-08-17 four-repository homepage routing",
+ "36,728 views / 8,881 unique visitors",
+ "4,580 / 1,499",
+ "8,315 / 2,996",
+ "3,408 / 1,520",
+ "2,597 / 724",
+ "575 / 231",
+ "181 / 87",
+ "96 / 43",
+ "[#3508](https://github.com/modelscope/FunASR/pull/3508)",
+ "`2cb615c42911972602cfaec739144d406b3282c5`",
+ "`20260817T012430Z`",
+ "`20260817T004144Z`",
+ "`5c5b96595d4cb893c83aedeca4d5dfe5a4cf0a665a1b4d489808886a7aacc1a3`",
+ "Production speech toolkit",
+ "High-accuracy ASR",
+ "Languages and emotion",
+ "AI video editing",
+ ]
+ for marker in required_markers:
+ assert marker in text