diff --git a/CLAUDE.md b/CLAUDE.md
index 82b72e5..220e3fa 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -141,7 +141,7 @@ When a PR merges, `action.py` runs this sequence:
    If true → exit 0, log "low-signal PR skipped"
 6. Call `extract(title, body, review_comments)` — Instructor + Claude Sonnet
 7. If `contains_decision` is False → exit 0 silently
-7b. If `contains_decision` is True but `confidence < 0.65` → discard silently
+7b. If `contains_decision` is True but `confidence < 0.40` → discard silently
 8. If `contains_decision` is True:
    - Scan PR body + review comments for `ADR-NNN` patterns; glob matching knowledge records
      into a `related` list
diff --git a/README.md b/README.md
index db02afe..284a059 100644
--- a/README.md
+++ b/README.md
@@ -1,8 +1,14 @@
 # Memex
 
-Memex extracts and indexes the institutional knowledge buried in your GitHub pull requests, review threads, and ADRs — automatically, without changing how your team works.
+[![PyPI](https://img.shields.io/pypi/v/memex-oss)](https://pypi.org/project/memex-oss/)
+[![Python 3.12+](https://img.shields.io/badge/python-3.12%2B-blue)](https://pypi.org/project/memex-oss/)
+[![License: AGPL v3](https://img.shields.io/badge/License-AGPL%20v3-blue.svg)](https://www.gnu.org/licenses/agpl-3.0)
 
-Engineers don't write anything new. Memex hooks into the artifacts that already exist and makes them searchable.
+**[trymemex.dev](https://trymemex.dev)** · [PyPI](https://pypi.org/project/memex-oss/)
+
+Most architectural decisions are never written down. They live in a three-minute hallway conversation, a Slack thread that gets deleted, or a pull request where every review is just "LGTM".
+
+Memex captures decisions at the moment they're made — extracting structured rationale from the GitHub artifacts that already exist (PRs, review threads, ADRs) without asking engineers to change how they work.
 
 ```bash
 $ memex query "why did we move off MongoDB"
@@ -19,6 +25,10 @@ Results for: why did we move off MongoDB
   #2  ...
 ```
 
+Or, surfaced automatically by an AI coding agent before it suggests an architecture change:
+
+![Cursor using memex_query to surface a past decision before suggesting an architecture change](memex_cursor_landing.html.png)
+
 ---
 
 ## How it works
diff --git a/memex/schema.py b/memex/schema.py
index 7edc227..aa745c5 100644
--- a/memex/schema.py
+++ b/memex/schema.py
@@ -6,7 +6,7 @@
 class ConfidenceLevel(str, Enum):
     high = "high"       # >= 0.80 — publish
     medium = "medium"   # 0.65–0.80 — publish with flag
-    low = "low"         # < 0.65 — discard
+    low = "low"         # < 0.65 — publish with ⚠️ flag
 
 
 class KnowledgeRecord(BaseModel):
diff --git a/memex_cursor_landing.html.png b/memex_cursor_landing.html.png
new file mode 100644
index 0000000..a3586d4
Binary files /dev/null and b/memex_cursor_landing.html.png differ
diff --git a/pyproject.toml b/pyproject.toml
index 2c97bc1..9266f42 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,12 +4,12 @@ build-backend = "hatchling.build"
 
 [project]
 name = "memex-oss"
-version = "0.3.0"
+version = "0.3.1"
 description = "Institutional knowledge extraction for engineering teams"
 readme = "README.md"
 license = { text = "AGPL-3.0-or-later" }
 authors = [{ name = "Memex Contributors" }]
-keywords = ["knowledge-management", "github-actions", "adr", "engineering"]
+keywords = ["knowledge-management", "github-actions", "adr", "engineering", "llm", "mcp", "developer-tools", "institutional-knowledge"]
 classifiers = [
     "Development Status :: 3 - Alpha",
     "Intended Audience :: Developers",
@@ -34,6 +34,12 @@ dev = [
     "pytest-mock>=3.14.0",
 ]
 
+[project.urls]
+Homepage = "https://trymemex.dev"
+Repository = "https://github.com/Jonasb8/memex"
+"Bug Tracker" = "https://github.com/Jonasb8/memex/issues"
+Changelog = "https://github.com/Jonasb8/memex/releases"
+
 [project.scripts]
 memex = "memex.cli:cli"