Alteran term

Sovereign inference

Running the model on hardware you control, with no cloud call in the loop: data, cost and failure modes all stay yours.

Related
10 articles
04 · SOVEREIGN INFERENCEA 26B model on 24 GB, without swapping

Gemma-4 26B swapped 27,848 times in a real agentic run. One flag reported zero, on an easy test. Under the load we actually run, it took two settings together: experts on the CPU and an 8-bit KV cache. No new runtime needed.

1 Nov 2026
06 · WORKING WITH CLAUDENothing to migrate

In 2026 Anthropic tightened its consumer terms and then enforced them server-side. Tools built on subscription tokens broke. Team Brain had nothing to change, not because anyone foresaw it, but because the shortcut was never taken.

1 Nov 2026
04 · SOVEREIGN INFERENCE24.2 seconds to 1.0: one flag in three places

Our local agent took 24 seconds to say 'OK'. The loop was fine. A cache flag was set to zero, in three places, and the one we edited first was not the live one.

1 Nov 2026
04 · SOVEREIGN INFERENCEOrder the prompt by volatility

On a paid API, prompt size is the enemy. On a local model with a prompt cache, volatility is. A 21,900-token per-turn bill, where the tool schemas were the bigger half, and the one ordering rule that serves both.

1 Nov 2026
04 · SOVEREIGN INFERENCERSS lied by four gigabytes

A plan blamed our local model for 13 GB of wired memory. Measured: its real footprint was 4.3 GB, its RSS said 0.19. Generation ran at a flat 11.5 tokens per second, and swap did not move during decode. A hypothesis of ours fell on the way.

1 Nov 2026
06 · WORKING WITH CLAUDESend the question, not the pixels

Remote computer use spent five full screenshots to learn that a window was black. One local call to our own vision model answered the question in a sentence, without the image leaving the machine. Local won on cost, speed, privacy and usefulness at once.

1 Nov 2026
05 · MULTI-AGENT, NOT A SWARM63 tool errors, and not one was the model reasoning badly

Our local agent logged 477 tool calls with 63 errors. Counted, it looks like an unreliable model. Classified by cause, 52 were our tooling, our guards or a deliberate test. The remaining 11 were aiming errors. 75% recovered on the next call.

1 Nov 2026
10 · PRODUCTSThree apps you can install

Crystal RAG, The Wisdomkeeper and SEO Brain: what each does, what it runs on, and why all three keep your data on your machine or your own server.

1 Nov 2026
04 · SOVEREIGN INFERENCE57 runs: the settings that did nothing

Context size, prompt-cache allocation, batch size: the standard advice for running a model locally. We tested it on one 12B model with tasks that had failed in production. None of it moved quality or speed. Bigger batches made it worse.

1 Nov 2026
04 · SOVEREIGN INFERENCEWhat we build

Sovereign AI systems: they run on hardware you own, and the data never leaves it. What that buys, what it costs, and the numbers.

1 Nov 2026

Glossary →