diff --git a/.changes/unreleased/Fixed and Improvements-20241126-151903.yaml b/.changes/unreleased/Fixed and Improvements-20241126-151903.yaml
deleted file mode 100644
index 2c1b320ac64b..000000000000
--- a/.changes/unreleased/Fixed and Improvements-20241126-151903.yaml	
+++ /dev/null
@@ -1,3 +0,0 @@
-kind: Fixed and Improvements
-body: Fixed a panic that occurred when specifying a local model ([#3464](https://github.com/TabbyML/tabby/issues/3464))
-time: 2024-11-26T15:19:03.863207+08:00
diff --git a/.changes/v0.21.0.md b/.changes/v0.21.0.md
new file mode 100644
index 000000000000..901f4d03e334
--- /dev/null
+++ b/.changes/v0.21.0.md
@@ -0,0 +1,19 @@
+## v0.21.0 (2024-12-02)
+
+### Notice
+
+* Due to changes in the indexing format, the `~/.tabby/index` directory will be automatically removed before any further indexing jobs are run. It is expected that the indexing jobs will be re-run (instead of incrementally) after the upgrade.
+
+### Features
+
+* Support connecting to llamafile model backend.
+* Display **Open** / **Closed** state for issues / pull requests in Answer Engine context card.
+* Support deleting the entire thread in Answer Engine.
+* Add rate limiter options for HTTP-powered model backends.
+
+### Fixed and Improvements
+
+* Fixed a panic that occurred when specifying a local model ([#3464](https://github.com/TabbyML/tabby/issues/3464))
+* Add pagination to Answer Engine threads.
+* Fix Vulkan binary distributions.
+* Improve the retry logic for chunk embedding computation in indexing job.
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 19e61b78463f..55c02fdcbd47 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,6 +5,26 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html),
 and is generated by [Changie](https://github.com/miniscruff/changie).
 
+## v0.21.0 (2024-12-02)
+
+### Notice
+
+* Due to changes in the indexing format, the `~/.tabby/index` directory will be automatically removed before any further indexing jobs are run. It is expected that the indexing jobs will be re-run (instead of incrementally) after the upgrade.
+
+### Features
+
+* Support connecting to llamafile model backend.
+* Display **Open** / **Closed** state for issues / pull requests in Answer Engine context card.
+* Support deleting the entire thread in Answer Engine.
+* Add rate limiter options for HTTP-powered model backends.
+
+### Fixed and Improvements
+
+* Fixed a panic that occurred when specifying a local model ([#3464](https://github.com/TabbyML/tabby/issues/3464))
+* Add pagination to Answer Engine threads.
+* Fix Vulkan binary distributions.
+* Improve the retry logic for chunk embedding computation in indexing job.
+
 ## v0.20.0 (2024-11-08)
 
 ### Features