oioxo-mcp 0.5.3 → 0.5.5

package/README.md

@@ -1,12 +1,12 @@
-# OIOXO — Context Engine + On-Device Coding Agent
+# OIOXO — cut your AI bill ~90%, and code privately from the terminal
 
-**Your AI agents read 10× more code than they need. You pay for every token of it.**
+**Your AI tools read far more of your code than they need — and you pay for every word.**
 
-OIOXO does two things from one CLI:
-1. **Context engine** — sits between your codebase and your AI agents (GitHub Copilot, Claude Code, Cursor, Windsurf, Gemini CLI, Codex) and hands them the *minimal relevant slice* of your project instead of letting them read whole files.
-2. **Coding agent** — `oioxo code "<task>"` plans, edits, and verifies in your repo, using your own key **or a local Ollama model** — nothing leaves your machine. See [Code in your terminal](#code-in-your-terminal--oioxo-code).
+OIOXO fixes that from one command-line tool. It does two things:
+1. **Makes Copilot, Cursor & Claude ~90% cheaper.** It sits between your project and your AI tools and hands them only the small, relevant slice — instead of letting them read whole files. Same answers, a fraction of the cost.
+2. **Codes for you, privately.** `oioxo code "<task>"` plans, edits, and checks the work right in your repo — using your own key **or a local model** — so nothing leaves your machine. See [Code in your terminal](#code-in-your-terminal--oioxo-code).
 
-**Measured on a real production codebase: 90–92% fewer context tokens per question.** A question that costs an agent ~50,000 tokens of file reading comes back as a ~5,000-token capsule — same answer, a fraction of the cost.
+**~90% fewer words per question** (measured on a real codebase). A question that would cost ~50,000 words comes back as ~5,000 — same answer, a fraction of the cost.
 
 ## Quick start
 
@@ -34,18 +34,26 @@ ANTHROPIC_API_KEY=… oioxo code "fix the failing test" --verify "npm test"
 OIOXO_PROVIDER=ollama OIOXO_MODEL=qwen2.5-coder oioxo code "add input validation to the signup form"
 ```
 
-- **Verified, not vibes** — every change is run through your tests/typecheck on a shadow copy; a red result is fed back and repaired. Your real files are only touched after you approve the diff.
+- **Verified, not vibes** — every change runs through your tests/typecheck on a shadow copy; a red result is fed back and repaired. Your real files are touched only after you approve the diff.
 - **Bring any model** — OpenAI, Anthropic, Groq, Mistral, Together, local Ollama, or any OpenAI-compatible base (`OIOXO_PROVIDER` / `OIOXO_BASE_URL` / `OIOXO_MODEL` / `OIOXO_API_KEY`).
-- **Or fully local, no key** — point it at [Ollama](https://ollama.com) (`OIOXO_PROVIDER=ollama OIOXO_MODEL=qwen2.5-coder`) and nothing leaves your machine. A built-in `--local` on-device coder (open model via llama.cpp, auto-downloaded once to `~/.oioxo/models`) is **coming next**.
+- **Or fully local, no key** — point it at [Ollama](https://ollama.com) (`OIOXO_PROVIDER=ollama OIOXO_MODEL=qwen2.5-coder`) and nothing leaves your machine. Or run the built-in on-device coder with **`oioxo code --local`** (open model via llama.cpp; one-time `npm i -g node-llama-cpp`) — no key, nothing uploaded.
 
-Flags: `--verify "<cmd>"` (override the test command) · `--yes` (apply without the prompt) · `--max-iters N`. *(`--local` is reserved for the built-in on-device coder, coming next — use `OIOXO_PROVIDER=ollama` for local today.)*
+Flags: `--verify "<cmd>"` (override the test command) · `--yes` (apply without the prompt) · `--max-iters N` · `--local` (built-in on-device coder, no key — one-time `npm i -g node-llama-cpp`).
+
+## Let your devices build together
+
+Put two of your own devices on the **same Wi-Fi** and they work as one — this machine can lend
+its power to another device's build (so your phone or a thin laptop builds on *this* machine's
+GPU), or borrow power the other way. Start it with `oioxo invite` (or `oioxo join`); pairing is
+just a short code (a QR or a string), with nothing going through our servers. Lend power and you
+earn **free coding time** for it. *(Feature name: Compute Mesh.)*
 
 ## Why developers use it
 
-- 💸 **Subscriptions go further** — Copilot premium requests, Claude limits, API keys all burn down slower.
-- ⚡ **Better answers** — focused context beats 75k tokens of noise; your agent gets the exact code in play and the parts of the project it actually depends on.
-- 🔒 **100% on-device** — your code is indexed and queried locally, never uploaded. Only the saved-token *count* is metered.
-- 🪄 **One command** — `init` detects Claude Code, VS Code/Copilot, Cursor, Windsurf, Gemini CLI and Codex, and merges their configs without touching your other MCP servers.
+- 💸 **Your AI costs drop ~90%** — Copilot premium requests, Claude limits, and API keys all last several times longer.
+- ⚡ **Better answers, too** — focused context beats 50k words of noise. Your AI gets the exact code in play, plus the parts it actually depends on.
+- 🔒 **100% on your device** — your code is read and searched locally, never uploaded. Only the *amount* you saved is counted.
+- 🪄 **One command** — `init` finds Claude Code, VS Code/Copilot, Cursor, Windsurf, Gemini CLI and Codex, and sets them up without touching your other tools.
 
 ## How it works (for you)
 
@@ -65,16 +73,40 @@ Flags: `--verify "<cmd>"` (override the test command) · `--yes` (apply without
 ## Commands
 
 ```
-oioxo-mcp code "<task>"    code in this repo: capsule-grounded, verified, then asks
-                           before writing (--verify "<cmd>" · --yes · --max-iters)
-oioxo-mcp login            connect your OIOXO account
-oioxo-mcp init [--all]     write agent MCP configs + instruction files
-oioxo-mcp serve            run the MCP server (agent configs call this)
-oioxo-mcp status           plan, savings and index stats
-oioxo-mcp logout           remove the stored credential
+# Code
+oioxo code "<task>"        code in this repo: capsule-grounded, verified, then asks
+                           before writing (--verify "<cmd>" · --yes · --max-iters · --local)
+                           --local runs the built-in on-device coder, no key
+
+# Account
+oioxo login | logout       connect / disconnect your OIOXO account
+oioxo whoami | usage        who you're signed in as · your saved-token allowance
+oioxo status               plan, savings and on-device index stats
+
+# Devices (Compute Mesh) — build together on the same Wi-Fi, one account
+oioxo invite               show a code; the device that HAS the project scans it,
+                           and this machine lends its coder to that build
+oioxo join                 scan/paste the code from the device that has the project,
+                           and lend this machine's compute to it
+oioxo devices              find your other same-account devices on this Wi-Fi
+oioxo stop-helping         how to stop lending (Ctrl+C in the helper's terminal)
+
+# AI / model
+oioxo model list | use <p> [m]    show providers / pin one (and a model)
+oioxo config set|get|list|unset   persist provider/model/baseUrl/apiKey
+
+# Agents (MCP)
+oioxo init [--all]         wire OIOXO into the AI agents in this project
+oioxo mcp list             show which agents OIOXO is wired into
+oioxo serve                run the MCP server over stdio (agents call this)
+
+# Maintenance
+oioxo doctor               check creds, connection, coder config, agent wiring
+oioxo update               how to update the CLI
+oioxo --version | --help
 ```
 
-> The binary is published as `oioxo-mcp`; `oioxo code …` and `oioxo-mcp code …` are the same.
+> The binary is published as `oioxo-mcp`; `oioxo …` and `oioxo-mcp …` are the same command.
 
 ## Pricing