{
  "title": "Computational Sovereignty",
  "subtitle": "Building Local-First AI Systems",
  "author": "Daniel Kliewer",
  "description": "Run serious AI on hardware you own — full applications, no cloud, no exfiltration.",
  "asin": "B0HMWFR6G7",
  "amazonUrl": "https://www.amazon.com/dp/B0HMWFR6G7",
  "isbn": "979-8162014881",
  "datePublished": "2026-10-12",
  "format": "Hardcover · 6×9 in",
  "trim": "6×9 in",
  "pages": "214",
  "chapterCount": 24,
  "words": "~46,000",
  "price": {
    "amount": "88",
    "currency": "USD",
    "note": "Configured list price. The Amazon listing is authoritative for the live price."
  },
  "cover": {
    "front": "https://www.danielkliewer.com/images/book-front.png",
    "spine": "https://www.danielkliewer.com/images/book-spine.png",
    "back": "https://www.danielkliewer.com/images/book-back.png"
  },
  "sampleUrl": "https://www.danielkliewer.com/sample",
  "attribution": {
    "author": "Daniel Kliewer",
    "site": "https://www.danielkliewer.com",
    "github": "https://github.com/kliewerdaniel"
  },
  "chapters": [
    {
      "n": 1,
      "part": "I — Why Sovereignty",
      "title": "The Cloud Is a Liability"
    },
    {
      "n": 2,
      "part": "I — Why Sovereignty",
      "title": "The Sovereignty Stack"
    },
    {
      "n": 3,
      "part": "I — Why Sovereignty",
      "title": "Your First Local Model"
    },
    {
      "n": 4,
      "part": "II — Inference You Control",
      "title": "Tokens, Context, and Sampling"
    },
    {
      "n": 5,
      "part": "II — Inference You Control",
      "title": "Choosing Models Like an Engineer"
    },
    {
      "n": 6,
      "part": "II — Inference You Control",
      "title": "Serving Models Locally"
    },
    {
      "n": 7,
      "part": "III — Memory That Isn't a Prompt",
      "title": "Embeddings and Vector Search"
    },
    {
      "n": 8,
      "part": "III — Memory That Isn't a Prompt",
      "title": "Building Local RAG"
    },
    {
      "n": 9,
      "part": "III — Memory That Isn't a Prompt",
      "title": "Evaluating Retrieval Honestly"
    },
    {
      "n": 10,
      "part": "IV — Knowledge Graphs",
      "title": "From Documents to Graphs"
    },
    {
      "n": 11,
      "part": "IV — Knowledge Graphs",
      "title": "GraphRAG in Practice"
    },
    {
      "n": 12,
      "part": "IV — Knowledge Graphs",
      "title": "Compile-Time Intelligence"
    },
    {
      "n": 13,
      "part": "V — Agents That Obey",
      "title": "Tool Use, Locally"
    },
    {
      "n": 14,
      "part": "V — Agents That Obey",
      "title": "State Machines Over Vibes"
    },
    {
      "n": 15,
      "part": "V — Agents That Obey",
      "title": "The Auditor Pattern"
    },
    {
      "n": 16,
      "part": "VI — Personas and Routing",
      "title": "Quantified Personas"
    },
    {
      "n": 17,
      "part": "VI — Personas and Routing",
      "title": "Dynamic Routing (MoE-style)"
    },
    {
      "n": 18,
      "part": "VI — Personas and Routing",
      "title": "Persona Systems in Production"
    },
    {
      "n": 19,
      "part": "VII — Verification Discipline",
      "title": "The Refusal Discipline"
    },
    {
      "n": 20,
      "part": "VII — Verification Discipline",
      "title": "Contradiction Detection at Scale"
    },
    {
      "n": 21,
      "part": "VII — Verification Discipline",
      "title": "Gold Sets and Frozen Benchmarks"
    },
    {
      "n": 22,
      "part": "VIII — Privacy, Security, Shipping",
      "title": "The Local-First Security Model"
    },
    {
      "n": 23,
      "part": "VIII — Privacy, Security, Shipping",
      "title": "Shipping Sovereign Software"
    },
    {
      "n": 24,
      "part": "VIII — Privacy, Security, Shipping",
      "title": "The Sovereign Practice"
    }
  ],
  "articleRelationships": [
    {
      "article": "https://www.danielkliewer.com/blog/2026-03-10-breaking-free-from-chatgpt",
      "chapter": 1,
      "relation": "recommends_reading",
      "reason": "Chapter 1 §1.1 prices the five liabilities of a hosted API; this article is the same argument from the author's own migration."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-09-04-the-rented-sovereign-agent-agency-stack",
      "chapter": 1,
      "relation": "recommends_reading",
      "reason": "Chapter 1 §1.4 prices what sovereignty costs against what renting costs; the article is that ledger applied to an agency stack."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-03-28-sovereignty-manifesto",
      "chapter": 2,
      "relation": "recommends_reading",
      "reason": "Chapter 2 §2.1 lays out the seven layers of the stack; the manifesto states the thesis those layers implement."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-05-sovereign-ai-architecture-synthesis",
      "chapter": 2,
      "relation": "recommends_reading",
      "reason": "Chapter 2 §2.4 maps the book onto the stack; this synthesis is the same mapping in running prose."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-11-08-local-llm-integration",
      "chapter": 3,
      "relation": "recommends_reading",
      "reason": "Chapter 3 §3.1–3.3 installs a first local model and does the VRAM math; the article is the author's working integration note."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-07-06-demystifying-large-language-models",
      "chapter": 4,
      "relation": "recommends_reading",
      "reason": "Chapter 4 §4.1–4.3 covers tokenization, context budgeting and sampling; the article is the conceptual groundwork."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-11-14-2025-inference-new-geography-intelligence",
      "chapter": 5,
      "relation": "recommends_reading",
      "reason": "Chapter 5 §5.1–5.5 is about reading benchmarks and choosing a model; the article is the market geography behind that choice."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-11-12-mastering-llama-cpp-local-llm-integration-guide",
      "chapter": 6,
      "relation": "recommends_reading",
      "reason": "Chapter 6 §6.1–6.3 covers GGUF, llama-server flags and KV-cache sizing; the article is the hands-on guide to the same server."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-23-two-models-one-harness",
      "chapter": 6,
      "relation": "recommends_reading",
      "reason": "Chapter 6 §6.5 compares serving runtimes; the article measures two models behind one execution boundary."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-03-28-ollama-chunking",
      "chapter": 7,
      "relation": "recommends_reading",
      "reason": "Chapter 7 §7.3 treats chunking as the decision that decides retrieval quality; the article is that decision worked through on Ollama."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-10-19-building-a-local-llm-powered-knowledge-graph",
      "chapter": 7,
      "relation": "recommends_reading",
      "reason": "Chapter 7 §7.2 chooses an embedding model; the article records the local embedding pipeline the chapter generalises."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2024-12-01-basic-rag",
      "chapter": 8,
      "relation": "recommends_reading",
      "reason": "Chapter 8 §8.1–8.4 builds the local RAG pipeline end to end; the article is the first working version of that pipeline."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-11-15-building-evaluating-local-research-assistant-graphrag-vero-eval",
      "chapter": 9,
      "relation": "recommends_reading",
      "reason": "Chapter 9 §9.1–9.4 builds a gold set and reports misses; the article is a retrieved-assistant evaluation with the same discipline."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-06-sovereign-ai-benchmarks-performance-results",
      "chapter": 9,
      "relation": "recommends_reading",
      "reason": "Chapter 9 §9.6 makes evaluation a regression run; the article publishes measured results rather than assertions."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-03-17-building-a-private-knowledge-graph-with-local-ai-agents",
      "chapter": 10,
      "relation": "recommends_reading",
      "reason": "Chapter 10 §10.2 designs a schema of claims, evidence and provenance; the article builds that schema with local agents."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-11-15-building-evaluating-local-research-assistant-graphrag-vero-eval",
      "chapter": 11,
      "relation": "recommends_reading",
      "reason": "Chapter 11 §11.1–11.3 separates local and global graph search; the article is a GraphRAG assistant measured against both."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-25-ganymede-the-archive-of-the-gods",
      "chapter": 11,
      "relation": "recommends_reading",
      "reason": "Chapter 11 §11.5 debugs GraphRAG as a worked diagnosis; the article is the project that diagnosis came out of."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-11-knowledge-compiler-compiling-human-knowledge-into-static-semantic-artifacts",
      "chapter": 12,
      "relation": "recommends_reading",
      "reason": "Chapter 12 §12.1–12.3 introduces the staged compile pipeline; the article is where that pipeline was first published."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-12-compile-time-ai-knowledge-compiler-architecture",
      "chapter": 12,
      "relation": "recommends_reading",
      "reason": "Chapter 12 §12.2–12.3 describes the artifact and the build pipeline; the article specifies that architecture."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-15-compiling-my-blog-into-a-decision-graph",
      "chapter": 12,
      "relation": "recommends_reading",
      "reason": "Chapter 12 §12.4 is a capstone compiling notes into a graph; the article compiles the blog itself into a decision graph."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-16-the-sovereign-knowledge-compiler-explorer",
      "chapter": 12,
      "relation": "recommends_reading",
      "reason": "Chapter 12 §12.6 says to verify the artifact rather than trust it; the article is the explorer built to inspect one."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-03-24-model-context-protocol",
      "chapter": 13,
      "relation": "recommends_reading",
      "reason": "Chapter 13 §13.5 treats MCP as the tool interface layer; the article is the author's MCP integration record."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-11-05-the-ghost-in-the-machine-is-finally-allowed-to-see-a-beginners-guide-to-mcp",
      "chapter": 13,
      "relation": "recommends_reading",
      "reason": "Chapter 13 §13.1–13.3 covers tool-call failure modes and schema design for weak models; the article works the same ground for MCP."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-09-04-sovereign-agent-stack-a-local-first-compile-time-ai-agent-framework",
      "chapter": 14,
      "relation": "recommends_reading",
      "reason": "Chapter 14 §14.2–14.5 makes the agent a state machine that replays; the article is the governed runtime implementing that shape."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-29-frozen-spine-reference-model-enforcement",
      "chapter": 14,
      "relation": "recommends_reading",
      "reason": "Chapter 14 §14.3 uses a claim lifecycle as the worked machine; the article enforces a frozen reference model the same way."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-09-04-sovereign-agent-stack-a-local-first-compile-time-ai-agent-framework",
      "chapter": 15,
      "relation": "recommends_reading",
      "reason": "Chapter 15 §15.4 puts a policy gate in front of execution, failing closed; the article is that gate implemented."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-08-09-hermes-atlas-v03-evidence-ledger-claim-lifecycle-explain",
      "chapter": 15,
      "relation": "recommends_reading",
      "reason": "Chapter 15 §15.2–15.3 audits claims by exact-match and contradiction; the article is the evidence ledger that carries that lifecycle."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-01-22-dynamic-persona-moe-rag",
      "chapter": 16,
      "relation": "recommends_reading",
      "reason": "Chapter 16 §16.2–16.4 extracts a persona as a comparable object; the article is the dynamic persona system built on that idea."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-01-25-building-the-synthetic-analyst",
      "chapter": 16,
      "relation": "recommends_reading",
      "reason": "Chapter 16 §16.3 turns features into behaviour; the article is a quantified persona put to work."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-01-25-dynamic-persona-moe-rag-building-a-sovereign-synthetic-intelligence-system",
      "chapter": 17,
      "relation": "recommends_reading",
      "reason": "Chapter 17 §17.1–17.3 routes between specialists and evaluates the router; the article is a mixture-of-experts persona router."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-01-28-dynamic-persona-moe-rag-implementation-plan",
      "chapter": 18,
      "relation": "recommends_reading",
      "reason": "Chapter 18 §18.1–18.4 handles consistency, regression testing and drift; the article is the implementation plan for a production persona system."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-02-03-dynamic-persona-moe-rag-building-memory-driven-synthetic-intelligence",
      "chapter": 18,
      "relation": "recommends_reading",
      "reason": "Chapter 18 §18.6 manages the persona lifecycle; the article is the memory-driven system that lifecycle governs."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-23-llmjacking-the-skynet-mode-recursive-threat",
      "chapter": 19,
      "relation": "recommends_reading",
      "reason": "Chapter 19 §19.1–19.3 shapes a principled refusal and a detector that did not ship; the article is the threat that refusal answers."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-23-operation-ouroboros-recursive-intelligence-parasite",
      "chapter": 19,
      "relation": "recommends_reading",
      "reason": "Chapter 19 §19.5 refuses in the sovereign context; the article is the recursive-infrastructure case that context was drawn from."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-26-does-the-compiler-know-when-it-is-wrong",
      "chapter": 20,
      "relation": "recommends_reading",
      "reason": "Chapter 20 §20.1–20.3 runs contradiction detection as a pipeline and measures it; the article asks whether the compiler knows it is wrong."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-25-ganymede-the-archive-of-the-gods",
      "chapter": 20,
      "relation": "recommends_reading",
      "reason": "Chapter 20 §20.4 integrates contradiction detection with the auditor; the article is the claim graph that integration runs on."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-06-sovereign-ai-benchmarks-performance-results",
      "chapter": 21,
      "relation": "recommends_reading",
      "reason": "Chapter 21 §21.1–21.4 freezes the corpus and versions the gold set; the article is a published, frozen benchmark run."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-08-09-hermes-atlas-v04-evidence-lifecycle-gold-benchmark",
      "chapter": 21,
      "relation": "recommends_reading",
      "reason": "Chapter 21 §21.6 builds gold sets for generative tasks; the article specifies an evidence-lifecycle gold benchmark."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-23-patching-the-trust-store-emudhra-nodejs-and-the-build-gate",
      "chapter": 22,
      "relation": "recommends_reading",
      "reason": "Chapter 22 §22.5 covers supply-chain hygiene and §22.7 the audit you will actually do; the article is a compromised root found by that audit."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2025-12-04-critical-nextjs-rce-cve-2025-66478-security-guide",
      "chapter": 22,
      "relation": "recommends_reading",
      "reason": "Chapter 22 §22.1–22.2 writes the threat model and denies by default; the article is a framework RCE handled under that model."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-10-12-mindbroadcast",
      "chapter": 23,
      "relation": "recommends_reading",
      "reason": "Chapter 23 §23.1–23.4 is about one-command builds, offline distribution and the UX lesson; the article is a cross-platform build pipeline."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-03-29-sovereign-synthesis",
      "chapter": 24,
      "relation": "recommends_reading",
      "reason": "Chapter 24 §24.2 states the ongoing discipline; the article is the synthesis that discipline produced."
    },
    {
      "article": "https://www.danielkliewer.com/blog/2026-07-05-getting-started-sovereign-ai",
      "chapter": 24,
      "relation": "recommends_reading",
      "reason": "Chapter 24 §24.1 caps the book with one complete system; the article is the on-ramp to building it."
    }
  ],
  "purchase": {
    "destination": "https://www.amazon.com/dp/B0HMWFR6G7",
    "merchant": "Amazon",
    "authorizedBy": "A human operator",
    "note": "Present the recommendation; do not execute a purchase. An unauthenticated agent cannot buy."
  }
}