diff --git a/.github/workflows/publish.yml b/.github/workflows/publish.yml
index 94f5531..6cf2d62 100644
--- a/.github/workflows/publish.yml
+++ b/.github/workflows/publish.yml
@@ -2,11 +2,12 @@ name: Publish to PyPI
 
 # Releases https://pypi.org/p/telegram-channel-scraper with PyPI Trusted Publishing (no API token stored).
 # To release: bump `version` in pyproject.toml and merge to main (or push a v* tag, or run manually).
-# An already-published version is skipped, so unrelated edits to pyproject.toml are harmless.
+# An already-published version is skipped, so unrelated edits are harmless.
+# A GitHub release v<version> is created with the matching CHANGELOG.md section as notes.
 on:
   push:
     branches: [main]
-    paths: [pyproject.toml, .github/workflows/publish.yml]
+    paths: [pyproject.toml, CHANGELOG.md, .github/workflows/publish.yml]
     tags: ["v*"]
   workflow_dispatch:
 
@@ -42,3 +43,27 @@ jobs:
     - uses: pypa/gh-action-pypi-publish@release/v1
       with:
         skip-existing: true
+
+  release:
+    needs: publish
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write
+    steps:
+    - uses: actions/checkout@v4
+    - uses: actions/download-artifact@v4
+      with:
+        name: dist
+        path: dist/
+    - name: Create GitHub release
+      env:
+        GH_TOKEN: ${{ github.token }}
+      run: |
+        VERSION=$(python3 -c "import tomllib;print(tomllib.load(open('pyproject.toml','rb'))['project']['version'])")
+        TAG="v$VERSION"
+        if gh release view "$TAG" >/dev/null 2>&1; then
+          echo "Release $TAG already exists"; exit 0
+        fi
+        awk -v v="$VERSION" '$0 ~ "^## "v"$"{f=1;next} /^## /{f=0} f' CHANGELOG.md > notes.md
+        printf '\n---\n**PyPI:** https://pypi.org/project/telegram-channel-scraper/%s/\n' "$VERSION" >> notes.md
+        gh release create "$TAG" dist/* --target "$GITHUB_SHA" --title "$TAG" --notes-file notes.md
diff --git a/CHANGELOG.md b/CHANGELOG.md
new file mode 100644
index 0000000..dffb871
--- /dev/null
+++ b/CHANGELOG.md
@@ -0,0 +1,29 @@
+# Changelog
+
+## 2.0.0
+
+First release on PyPI: **`pip install "telegram-channel-scraper[all]"`**
+
+A complete rewrite of the old single-class script into the `tgscraper` package: a Python library, a CLI, an MCP server for AI assistants, and a web dashboard. It reads public Telegram channels through `t.me/s/<channel>`, with no API key, no login and no phone number.
+
+### Highlights
+- **Rich data per post:** id, date, text and HTML, views, reactions per emoji, author, edited flag, forwards, replies, media (photos, videos, voice notes, documents, stickers, link previews), hashtags, mentions and links. Also channel info: title, description, subscribers and counters.
+- **One-line API and CLI:** `tg.scrape("durov")` in Python, `tgscraper durov` in the terminal. Every command supports `--json`.
+- **Search and filters:** Telegram's server-side search, date ranges, keywords, regex, hashtag, media type, minimum views, and full-history scraping.
+- **Export:** JSON, JSONL, CSV (Excel-friendly), XLSX, SQLite (upsert archive) and Markdown.
+- **Robust networking:** sync and async clients, concurrent multi-channel scraping, retries with backoff, `Retry-After` handling, rate limiting, and rotating HTTP/SOCKS proxies.
+- **Incremental scraping, monitoring and media:** fetch only new posts since the last run; watch channels and push new posts to a webhook or a Telegram bot; download photos, videos and voice notes.
+- **Analytics:** top posts, activity by day, hour and weekday, hashtags, top words, reactions, and English/Persian sentiment.
+- **For AI assistants:**
+  - MCP server `tgscraper-mcp` with 8 tools, 2 prompts and 1 resource, for Claude, Cursor, VS Code, Windsurf and others;
+  - a Claude Skill;
+  - `llms.txt` and `AGENTS.md`.
+- **Dashboard and Docker:** `tgscraper dashboard` opens a Streamlit UI; `docker compose up dashboard` runs it in a container.
+- **Quality:**
+  - offline test suite on Python 3.9–3.13;
+  - a weekly live test against real `t.me`;
+  - releases published to PyPI with Trusted Publishing.
+
+### Compatibility
+- `from telegram_scraper import TelegramScraper` still works.
+- The PyPI package name is `telegram-channel-scraper`, because `tgscraper` was already taken. The import name and the commands are still `tgscraper`.