@cachly-dev/mcp-server 0.10.75 → 0.10.77

package/README.md

@@ -331,6 +331,40 @@ sessions + WIP registry — typically one tool call.
 
 ---
 
+## Editor support matrix
+
+`npx @cachly-dev/mcp-server@latest setup` auto-detects and configures all of the
+following. Manual snippets are in the **Manual Setup** section below.
+
+| Editor / Client | Auto-setup | Config file written | Global config | Notes |
+|---|---|---|---|---|
+| **Claude Code** | ✅ | `~/.claude/mcp.json` + `.mcp.json` | ✅ global always | Runtime device-flow sign-in on first tool call |
+| **Cursor** | ✅ detected via `.cursor/` | `.cursor/mcp.json` | — | Project-level; restart Cursor after setup |
+| **Windsurf** | ✅ detected via `.windsurf/` | `.windsurf/mcp.json` | — | Project-level; restart Windsurf after setup |
+| **VS Code + Copilot** | ✅ detected via `.vscode/` | `.vscode/mcp.json` | — | Requires VS Code MCP extension or Copilot chat |
+| **Cline** | ✅ detected via VS Code | `.vscode/mcp.json` | — | Shares config with Copilot; restart VS Code |
+| **Continue.dev** | ✅ detected via `.continue/` | `.continue/config.json` | — | Uses `modelContextProtocolServers` key |
+| **Zed** | ✅ detected via `.zed/` | `.zed/settings.json` | — | Uses `context_servers` key |
+| **Windsurf (global)** | `setup --editor windsurf` | `~/.windsurf/mcp.json` | ✅ | Pass `--editor` to target global config |
+| **Any other MCP client** | `init --editor claude` | `.mcp.json` | — | Standard `mcpServers` stdio format |
+
+**Which sign-in path each editor uses:**
+
+| Scenario | Path |
+|---|---|
+| `setup` from a real terminal (TTY) | OAuth device-flow → browser click → API key saved automatically |
+| `setup` from VSCode task / CI (non-TTY) | Auto-detects non-interactive, opens browser with step-by-step guide, prints `CACHLY_JWT=... setup` instruction |
+| First tool call from Claude Code (no JWT yet) | Inline device-flow: MCP returns URL + code, browser opens automatically, next call proceeds |
+| `CACHLY_JWT=cky_live_xxx npx ... setup` | Skips auth step entirely, uses provided key |
+
+> **Tip — fastest per-project setup from inside Claude Code:**
+> ```
+> Set up cachly for this project: npx @cachly-dev/mcp-server@latest setup
+> ```
+> Claude runs it and restarts automatically.
+
+---
+
 ## Manual Setup
 
 <details>
@@ -384,6 +418,46 @@ On the first tool call your AI will prompt you to sign in — takes 10 seconds.
 
 ---
 
+## Self-hosting & BYOK
+
+cachly is **bring-your-own-key and self-host friendly out of the box** — no
+enterprise contract required to keep data in your own infra.
+
+**Bring your own embedding key (BYOK).** Semantic search runs on the embedding
+provider *you* choose. Set one env var and cachly auto-detects it; no key needed if
+you prefer cachly's server-side embeddings (uses your JWT):
+
+| Provider | Env var | Model |
+|---|---|---|
+| OpenAI | `OPENAI_API_KEY` | `text-embedding-3-small` |
+| Google Gemini | `GEMINI_API_KEY` | `text-embedding-004` |
+| Mistral | `MISTRAL_API_KEY` | `mistral-embed` |
+| Cohere | `COHERE_API_KEY` | `embed-english-v3.0` |
+| Ollama (local, free) | `OLLAMA_BASE_URL` | `nomic-embed-text` |
+| cachly (server-side) | *(none — uses JWT)* | managed |
+
+Force a specific one with `CACHLY_EMBED_PROVIDER=openai`. Run
+`npx @cachly-dev/mcp-server@latest health` to confirm which provider is active.
+
+**Point at your own backend (self-hosting).** Every cachly install can talk to a
+private backend instead of `api.cachly.dev`:
+
+```bash
+# One-shot: wire up the wizard against your self-hosted backend
+npx @cachly-dev/mcp-server@latest setup --api-url https://cachly.mycorp.internal
+
+# Or non-interactively
+npx @cachly-dev/mcp-server@latest init \
+  --instance-id <uuid> --api-key <cky_live_...> \
+  --api-url https://cachly.mycorp.internal
+```
+
+`setup`/`init` bake `CACHLY_API_URL` into the editor config **only** when it differs
+from the default cloud — so default installs stay clean, and self-hosted installs
+keep talking to your backend on every editor launch. All data stays in your infra.
+
+---
+
 ## Pricing
 
 | Tier | RAM | Price | Best for |

package/dist/src/handlers/brain.js

@@ -7,8 +7,22 @@ import { keywordSearch, tokenize, splitMultiQuery, levenshtein, indexVocab as _i
 import { rerankByQuality } from '../rerank.js';
 import { computeEmbedding, hasEmbedProvider } from '../embeddings.js';
 // ── Changelog (shown once per version in session_start) ──────────────────────
-const MCP_VERSION = '0.10.75';
+const MCP_VERSION = '0.10.77';
 const WHATS_NEW = {
+    '0.10.77': [
+        `🩹 **Frictionless onboarding + first-class self-hosting/BYOK**`,
+        `  🚀 Editor-launched server now serves all tools even before sign-in — zero-credential device flow finally triggers on first tool call`,
+        `  🔒 No more stdout pollution: stdio JSON-RPC stream stays clean when no JWT is set`,
+        `  🏠 \`setup --api-url\` / \`init --api-url\` point at a self-hosted backend; URL baked into config only when non-default`,
+        `  🧬 \`health\` now shows your BYOK embedding provider + accepts cky_ long-lived keys (was a false "invalid token" failure)`,
+    ],
+    '0.10.76': [
+        `📊 **Agent-trace benchmark + editor matrix + setup timer**`,
+        `  📂 New \`agent-traces-corpus.json\` — 22 lessons, 15 queries, realistic AI-agent debug patterns`,
+        `  🏆 Benchmark result: **P@1 +66.7%, MRR +9.5% vs flat-file** (adversarial failure-distractors per category)`,
+        `  📋 README: first-class editor support matrix (Claude Code, Cursor, Windsurf, VSCode, Cline, Continue.dev, Zed)`,
+        `  ⏱️  setup now shows elapsed time + "Brain is ready" banner at the end`,
+    ],
     '0.10.75': [
         `🩹 **Setup hardened — works from the VSCode plugin & any non-TTY terminal**`,
         `  🔌 \`setup\` no longer hangs on a non-interactive stdin (VSCode tasks, CI) — auto-detects + runs in automatic mode`,