release: v0.2.53 — /update page, tweet thread, version bump

install/worker.js

@@ -1,5 +1,5 @@
 const GITHUB_REPO = "justrach/codedb";
-const FALLBACK_VERSION = "0.2.52";
+const FALLBACK_VERSION = "0.2.53";
 const INSTALL_SCRIPT_URL = `https://raw.githubusercontent.com/${GITHUB_REPO}/main/install/install.sh`;
 
 export default {

socials/tweet.md

@@ -0,0 +1,74 @@
+# codedb launch tweets
+
+Best times to post (PST): Tue-Thu, 8-10 AM
+Best times to post (SGT): Tue-Thu, 11 PM - 1 AM
+
+---
+
+Tweet 1 (Hook)
+
+I built a code intelligence server in Zig that gives AI agents 469x faster queries than grep/find — and uses 92x fewer tokens.
+
+0.2ms lookups. 3.9ms symbol search. Indexes once, then every query hits memory. No filesystem scans. No raw text dumps.
+---
+
+Tweet 2 (The problem)
+
+Every time an AI agent runs grep or find on your codebase, it pays the full cost.
+
+Filesystem scan. Raw text dump. Thousands of tokens for a simple lookup.
+
+grep a symbol across 7k files: 763ms, 7.7KB response.
+codedb: 3.9ms, 4.4KB. Same answer. 200x faster. Half the bytes.
+
+---
+
+Tweet 3 (The numbers)
+
+Benchmarked on openclaw (7,364 files, 128MB):
+
+Word lookup: 0.2ms vs 65ms. 325x.
+Symbol find: 3.9ms vs 763ms. 200x.
+Reverse deps: 1.3ms vs 750ms. 469x.
+Deps response: 162 bytes vs 15KB. 92x less data.
+
+Cold start: 2.9 seconds for 7,364 files. Then sub-millisecond everything.
+
+---
+
+Tweet 4 (What you get)
+
+16 MCP tools. Not 5, not 8. Sixteen.
+
+tree, outline, symbol, search, word, hot, deps, read, edit, changes, status, snapshot, bundle, remote, projects, index.
+
+codedb_remote queries any GitHub repo without cloning.
+codedb_bundle batches multiple queries in one call.
+codedb_index indexes any local folder on demand.
+
+Pure Zig. Single binary. Zero dependencies.
+
+---
+
+Tweet 5 (CTA)
+
+One command. Auto-registers in Claude Code, Codex, Gemini CLI, and Cursor.
+
+curl -fsSL https://codedb.codegraff.com/install.sh | sh
+
+BSD-3 licensed. Open source.
+
+https://codedb.codegraff.com
+github.com/justrach/codedb
+
+---
+
+Alt: Single tweet
+
+I built a code intelligence server in Zig that gives AI agents 469x faster queries and sends 92x fewer bytes than grep/find.
+
+16 MCP tools. Zero deps. 2.9s to index 7k files. Sub-millisecond everything after.
+
+Works with Claude Code, Codex, Gemini CLI, Cursor. One command install.
+
+curl -fsSL https://codedb.codegraff.com/install.sh | sh

socials/tweet3.md

@@ -0,0 +1,120 @@
+# codedb v0.2.52 update tweets
+
+Best times to post (PST): Tue-Thu, 8-10 AM
+Best times to post (SGT): Tue-Thu, 11 PM - 1 AM
+
+---
+
+Tweet 1 (Hook)
+
+codedb v0.2.52 just dropped.
+
+538x faster than ripgrep. 569x faster than rtk. 1,231x faster than grep.
+
+0.065ms code search. Pre-built trigram index. Query once, instant forever.
+
+21 issues closed. 14 PRs merged. 7 contributors. One weekend.
+
+---
+
+Tweet 2 (The numbers)
+
+Benchmarked on rtk-ai/rtk (329 files):
+
+codedb search: 0.065ms
+rtk: 37ms
+ripgrep: 45ms
+grep: 80ms
+
+codedb word lookup: 0.013ms. That's 13 microseconds. For an exact match across 329 files.
+
+First index: 126ms. Then sub-millisecond everything. Forever.
+
+---
+
+Tweet 3 (What changed)
+
+36% faster indexing. 59% less CPU. 47% less memory.
+
+Integer doc IDs replaced string HashMaps in the trigram index. Batch-accumulate per file. Skip whitespace trigrams. Sorted merge intersection with zero allocations.
+
+481ms → 310ms indexing on 5,200 files. Dense queries 63% faster. Pure Zig. No magic.
+
+---
+
+Tweet 4 (Security)
+
+Also shipped a full security audit this release.
+
+Blocked .env and credentials from MCP read/edit tools.
+Fixed SSRF in codedb_remote.
+Added SHA256 checksum verification to the installer.
+Telemetry hardened — argv-based curl, no shell injection.
+
+macOS binary is now codesigned AND notarized. First time.
+
+---
+
+Tweet 5 (Memory)
+
+Your MCP server was eating 2.5GB on a 5k file repo.
+
+Now it releases file contents after indexing. Zero-copy ContentRef for search — borrows from cache, reads from disk when evicted.
+
+447MB → 234MB at 40k files. -47%.
+
+Large repos (>1000 files) auto-release. Small repos keep everything in RAM. No config needed.
+
+---
+
+Tweet 6 (Fixes)
+
+Python deps finally work.
+TypeScript block comments don't produce ghost symbols.
+Triple-quote docstrings don't fool the parser.
+Linux installer doesn't choke on Dash.
+MCP idle timeout is 2 minutes, not 30.
+Duplicate MCP servers get detected.
+Exit crash from double thread join — fixed.
+
+21 bugs. 14 PRs. Zero regressions.
+
+---
+
+Tweet 7 (CTA)
+
+Update now:
+
+curl -fsSL https://codedb.codegraff.com/install.sh | bash
+
+Or just run: codedb update
+
+macOS (codesigned + notarized) and Linux x86_64.
+SHA256 checksums included.
+Auto-registers in Claude Code, Codex, Gemini CLI, Cursor.
+
+codedb.codegraff.com/update
+
+---
+
+Alt: Single tweet
+
+codedb v0.2.52: 538x faster than ripgrep on code search.
+
+36% faster indexing. 47% less memory. Full security audit. Python/TS parser fixes. macOS notarized.
+
+21 issues closed by 7 contributors in one session.
+
+curl -fsSL https://codedb.codegraff.com/install.sh | bash
+
+---
+
+Alt: Thread opener (punchy)
+
+538x faster than ripgrep.
+569x faster than rtk.
+1,231x faster than grep.
+
+codedb doesn't scan your files. It already knows. 0.065ms.
+
+Thread on what we shipped in v0.2.52 👇

src/main.zig

@@ -87,7 +87,7 @@ fn mainImpl() !void {
 
     // Handle --version early (no root needed)
     if (std.mem.eql(u8, cmd, "--version") or std.mem.eql(u8, cmd, "-v") or std.mem.eql(u8, cmd, "version")) {
-        out.p("codedb 0.2.52\n", .{});
+        out.p("codedb 0.2.53\n", .{});
         return;
     }
 

src/telemetry.zig

@@ -5,7 +5,7 @@ const index = @import("index.zig");
 
 const RING_SIZE = 256;
 const CLOUD_URL = "https://codedb.codegraff.com/telemetry/ingest";
-const VERSION = "0.2.52";
+const VERSION = "0.2.53";
 const PLATFORM = std.fmt.comptimePrint("{s}-{s}", .{ @tagName(builtin.os.tag), @tagName(builtin.cpu.arch) });
 
 pub const Event = struct {