<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Arbiter Blog</title><description>Engineering notes from building Arbiter: orchestration, crash recovery, model routing, and the agnostic dispatch layer for your AI operation.</description><link>https://blog.arbitermcp.com/</link><language>en-us</language><item><title>One brain on your card, shared by the whole stack</title><link>https://blog.arbitermcp.com/one-brain-shared-by-the-whole-stack/</link><guid isPermaLink="true">https://blog.arbitermcp.com/one-brain-shared-by-the-whole-stack/</guid><description>Every local-first tool wants its own model on your GPU, and a single card runs out of room to hold them all. Arbiter keeps one resident brain and lets the rest of the stack borrow it, so Curator&apos;s memory work and Arbiter&apos;s own routing run on a single loaded model instead of a pile of duplicates.</description><pubDate>Fri, 21 Aug 2026 00:00:00 GMT</pubDate><category>local-llm</category><category>orchestration</category><category>ecosystem</category><author>Arbiter</author></item><item><title>Take the wheel of your own agent sessions, from anywhere, behind a second factor</title><link>https://blog.arbitermcp.com/drive-a-live-session-behind-a-second-factor/</link><guid isPermaLink="true">https://blog.arbitermcp.com/drive-a-live-session-behind-a-second-factor/</guid><description>Arbiter can drive your own live coding sessions remotely, including running work on them. Reading stays inside your own private bot channel, and every acting step is gated further by a single-use code from your authenticator. Here is how that stays safe.</description><pubDate>Thu, 20 Aug 2026 00:00:00 GMT</pubDate><category>remote-control</category><category>security</category><category>2fa</category><author>Arbiter</author></item><item><title>Arbiter&apos;s local brain just got more room to think</title><link>https://blog.arbitermcp.com/a-bigger-faster-local-brain/</link><guid isPermaLink="true">https://blog.arbitermcp.com/a-bigger-faster-local-brain/</guid><description>The on-device model that runs Arbiter&apos;s orchestration was upgraded. It now holds more of your operation at once and returns results in a shape you can rely on, all without a cloud call in the decision path.</description><pubDate>Sat, 15 Aug 2026 00:00:00 GMT</pubDate><category>local-llm</category><category>orchestration</category><category>performance</category><author>Arbiter</author></item><item><title>Supervise a fleet of coding agents, not a single window</title><link>https://blog.arbitermcp.com/supervise-a-fleet-of-coding-agents/</link><guid isPermaLink="true">https://blog.arbitermcp.com/supervise-a-fleet-of-coding-agents/</guid><description>One agent in one window is easy. A fleet of them is where the real work is, and where the tools stop helping. Arbiter launches, names, and drives many live sessions at once, with status you can trust about which ones are actually alive.</description><pubDate>Sun, 09 Aug 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>fleet</category><category>supervision</category><author>Arbiter</author></item><item><title>Route every task to the cheapest model that can actually do it</title><link>https://blog.arbitermcp.com/the-cheapest-sufficient-model/</link><guid isPermaLink="true">https://blog.arbitermcp.com/the-cheapest-sufficient-model/</guid><description>Frontier models get handed trivial work every day, and the bill never notices. Arbiter routes each task against the capability, cost, and safety policy you set, and absorbs the coordination work on a local brain that costs nothing per call.</description><pubDate>Tue, 04 Aug 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>cost</category><category>local-llm</category><author>Arbiter</author></item><item><title>Own the terminal. Don&apos;t live inside it.</title><link>https://blog.arbitermcp.com/own-the-terminal-as-a-system-layer/</link><guid isPermaLink="true">https://blog.arbitermcp.com/own-the-terminal-as-a-system-layer/</guid><description>Most AI coding work happens inside one terminal window at a time. Arbiter treats the terminal as a system layer it owns and drives from above, which is what lets one operator run a fleet instead of babysitting tabs.</description><pubDate>Mon, 27 Jul 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>positioning</category><author>Arbiter</author></item><item><title>Cheap by Default, Controlled From Your Phone</title><link>https://blog.arbitermcp.com/cheap-by-default-controlled-by-phone/</link><guid isPermaLink="true">https://blog.arbitermcp.com/cheap-by-default-controlled-by-phone/</guid><description>Make capability, cost, and safety tradeoffs explicit in routing policy, then use configured remote approval for sensitive work. Arbiter is a local-first working alpha in public binary release preparation.</description><pubDate>Mon, 20 Jul 2026 00:00:00 GMT</pubDate><category>ecosystem</category><category>cost</category><category>control</category><author>Arbiter</author></item><item><title>Approve Your Agent&apos;s Plan From Your Phone, Then Let It Run</title><link>https://blog.arbitermcp.com/approve-plans-from-your-phone/</link><guid isPermaLink="true">https://blog.arbitermcp.com/approve-plans-from-your-phone/</guid><description>With a supported remote channel and approval policy configured, review an agent plan from your phone, approve sensitive execution with a single-use code, and follow the session.</description><pubDate>Sun, 19 Jul 2026 00:00:00 GMT</pubDate><category>remote-control</category><category>orchestration</category><category>2fa</category><author>Arbiter</author></item><item><title>Guardrails for Autonomous Agents</title><link>https://blog.arbitermcp.com/guardrails-for-autonomous-agents/</link><guid isPermaLink="true">https://blog.arbitermcp.com/guardrails-for-autonomous-agents/</guid><description>Fanning an agent out across many workers is powerful and risky. Arbiter provides configurable controls for stopping work, limiting depth, setting budgets, and requiring approval, while interrupted side-effecting work remains visible for review.</description><pubDate>Mon, 13 Jul 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>safety</category><category>autonomy</category><author>Arbiter</author></item><item><title>What Is Remote Agent Supervision?</title><link>https://blog.arbitermcp.com/what-is-remote-agent-supervision/</link><guid isPermaLink="true">https://blog.arbitermcp.com/what-is-remote-agent-supervision/</guid><description>As developers hand more work to long-running AI agents, a new problem appears: supervising agents you cannot watch, from wherever you are, without shipping your keys or your code to someone else&apos;s cloud. Here is what good looks like.</description><pubDate>Tue, 07 Jul 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>supervision</category><category>security</category><author>Arbiter</author></item><item><title>Headless follow-ups into Claude and Codex</title><link>https://blog.arbitermcp.com/headless-followups-claude-and-codex/</link><guid isPermaLink="true">https://blog.arbitermcp.com/headless-followups-claude-and-codex/</guid><description>With a supported remote channel configured, send a running Claude or Codex session more work without opening a window or losing its context. Arbiter reaches the session it already has.</description><pubDate>Wed, 01 Jul 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>delegation</category><author>Arbiter</author></item><item><title>Remote control, locked down</title><link>https://blog.arbitermcp.com/remote-control-locked-down/</link><guid isPermaLink="true">https://blog.arbitermcp.com/remote-control-locked-down/</guid><description>Arbiter lets you drive your AI operation from your phone. That power needs a lock on it, so remote control sits behind a per-conversation code from your own authenticator.</description><pubDate>Sun, 28 Jun 2026 00:00:00 GMT</pubDate><category>security</category><category>remote-control</category><category>2fa</category><author>Arbiter</author></item><item><title>Your standup, run by a local brain</title><link>https://blog.arbitermcp.com/your-standup-run-by-a-local-brain/</link><guid isPermaLink="true">https://blog.arbitermcp.com/your-standup-run-by-a-local-brain/</guid><description>Arbiter&apos;s local brain can support the operations loop: a morning briefing across projects, a tool-calling chat, and configured dispatch to coding agents with policy-controlled approvals and stopping controls.</description><pubDate>Wed, 24 Jun 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>local-llm</category><category>supervision</category><author>Arbiter</author></item><item><title>Arbiter is alive: the orchestration layer, now running</title><link>https://blog.arbitermcp.com/arbiter-is-alive/</link><guid isPermaLink="true">https://blog.arbitermcp.com/arbiter-is-alive/</guid><description>Arbiter is a working proprietary alpha: a local brain on your own GPU, configured multi-provider routing, durable interruption handling, and configurable remote control. Public binary release preparation is underway, and there is an 8-minute walkthrough.</description><pubDate>Tue, 16 Jun 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>alpha</category><category>demo</category><author>Arbiter</author></item><item><title>Arbiter&apos;s brain runs on your GPU</title><link>https://blog.arbitermcp.com/the-brain-runs-on-your-gpu/</link><guid isPermaLink="true">https://blog.arbitermcp.com/the-brain-runs-on-your-gpu/</guid><description>Provider-agnostic should not stop at the providers. Arbiter can use a local model on your own hardware for routing assistance, without requiring a cloud call in that decision path.</description><pubDate>Thu, 11 Jun 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>local-llm</category><category>privacy</category><author>Arbiter</author></item><item><title>You already have an AI fleet. You just don&apos;t have a commander.</title><link>https://blog.arbitermcp.com/the-ai-stack-arbiter-orchestrates/</link><guid isPermaLink="true">https://blog.arbitermcp.com/the-ai-stack-arbiter-orchestrates/</guid><description>Arbiter is a local-first orchestration layer for configured agents, providers, and jobs. It can apply routing policy, preserve durable run state, and use a local GPU model for orchestration assistance.</description><pubDate>Wed, 10 Jun 2026 00:00:00 GMT</pubDate><category>ecosystem</category><category>orchestration</category><author>Arbiter</author></item><item><title>Provider-agnostic orchestration: why we don&apos;t lock you in</title><link>https://blog.arbitermcp.com/provider-agnostic-orchestration/</link><guid isPermaLink="true">https://blog.arbitermcp.com/provider-agnostic-orchestration/</guid><description>Why an orchestration layer should stay neutral about the models, clients, and accounts you bring to it.</description><pubDate>Mon, 08 Jun 2026 00:00:00 GMT</pubDate><category>orchestration</category><category>positioning</category><author>Arbiter</author></item><item><title>Introducing Arbiter: the orchestration layer your AI stack is missing</title><link>https://blog.arbitermcp.com/introducing-arbiter/</link><guid isPermaLink="true">https://blog.arbitermcp.com/introducing-arbiter/</guid><description>A provider-agnostic command center for configured LLM clients, terminals, and long-running jobs. It applies routing policy and preserves interrupted run state for deliberate recovery.</description><pubDate>Tue, 02 Jun 2026 00:00:00 GMT</pubDate><category>announcement</category><author>Arbiter</author></item></channel></rss>