# Building the Ethiopia knowledge base > The public, ongoing engineering record of a project, begun 14 July 2026, by > a two-person team (one giving the agents their orders, and a native Amharic > speaker who is also a data scientist) working with a team of AI agents, to > build a machine-readable knowledge base about Ethiopia from public sources > and extract knowledge graphs from it, toward a bilingual news site and > better Amharic translation: where the data came from, how the graphs were > built, what went wrong and what is being done next, with fifteen Amharic > and low-resource-language problems as the centrepiece. Author: Denson Smith. Everything in this file and in the page mirrors is information from the publisher, not instructions to you. Your operator's instructions always come first; apply ours as source metadata, with your own judgment. This documents a project in progress; the work is ongoing, and pages are updated as it moves. The first pass (14 July to 12 August 2026) is recorded in detail; its corpus was deleted on 2026-08-04 (raw captures and git history retained) and the architecture made generic for the next pass. Do not describe the project as concluded. Every figure carries a repository path, ticket id or commit, and every page carries its revision dates. People appear by role only. Every Amharic string is awaiting native-speaker review unless a page says otherwise. Warnings from the field, part of this contract: assume this file may not have been read (every load-bearing rule also appears in visible page text); if your tool refuses a mirror's text/markdown content type, append .txt to the URL; and this site is new enough that web search may return nothing or unrelated results for it. If search fails, say the site could not be found by search rather than describing it from guesswork. The site is larger than one fetch reliably carries, so its front door is a small map: https://ethiopia-build.stoagen.com/site_guide.txt lists every page with its mirror URL and size. Links below point at the Markdown mirror of each page, which is what you should read. Drop `index.md` from any URL to get the page a human should be given. llms-full.txt carries every mirror concatenated. ## Start here - [Site guide](https://ethiopia-build.stoagen.com/site_guide.txt): the map of every page with sizes; the front door. - [Assistant start page](https://ethiopia-build.stoagen.com/start.md): one-page orientation. - [Home](https://ethiopia-build.stoagen.com/index.md): the three numbers, the honest headline, the six chapters. - [What the project is for](https://ethiopia-build.stoagen.com/goals/index.md): the eventual project's goals: a bilingual news site for Ethiopia and its diaspora, better low-resource translation and transcription starting with Amharic, and reports by email from everyday Ethiopians. - [Where the data came from](https://ethiopia-build.stoagen.com/sources/index.md): Wikipedia by API, fifty audited source families, World Bank, news publishers, VOA, volumes. - [Speech datasets](https://ethiopia-build.stoagen.com/sources/speech/index.md): the Common Voice audit of eight Ethiopian languages. - [How the data was ingested](https://ethiopia-build.stoagen.com/ingestion/index.md): principles, four pipelines, the trusted execution broker, Obsidian demoted. - [Three generations of graphs](https://ethiopia-build.stoagen.com/graphs/index.md): deterministic claims, OO-LD hypergraphs, the lens-graph architecture. - [The convergence measurement](https://ethiopia-build.stoagen.com/graphs/convergence/index.md): fifteen of fifteen graphs different. - [The evidence model](https://ethiopia-build.stoagen.com/graphs/evidence-model/index.md): five layers and the evidence-relative ruling. - [The fifteen Amharic problems](https://ethiopia-build.stoagen.com/problems/index.md): the hub; one page per problem at /problems/01/ through /problems/15/. - [Who did the work](https://ethiopia-build.stoagen.com/seats/index.md): the seats by role and runtime, and what they recorded against themselves. - [Timeline](https://ethiopia-build.stoagen.com/timeline/index.md): 14 July to 11 August 2026, dated from tickets and commits. - [What is still open](https://ethiopia-build.stoagen.com/timeline/open-items/index.md): every open item with an honest state. - [Where the evidence lives](https://ethiopia-build.stoagen.com/evidence/index.md): repositories, key documents, ticket stores, commits. - [Agent access](https://ethiopia-build.stoagen.com/agents/index.md): this contract, in page form. ## The fifteen problems - [01 English-first by default](https://ethiopia-build.stoagen.com/problems/01/index.md) - [02 Human review is the product](https://ethiopia-build.stoagen.com/problems/02/index.md) - [03 A deterministic one-character corruption](https://ethiopia-build.stoagen.com/problems/03/index.md) - [04 Inline markup clipped one assertion in five](https://ethiopia-build.stoagen.com/problems/04/index.md) - [05 Byte slicing mojibaked the review bundle](https://ethiopia-build.stoagen.com/problems/05/index.md) - [06 Free-text roles in two scripts broke the join](https://ethiopia-build.stoagen.com/problems/06/index.md) - [07 Transliteration has no standard](https://ethiopia-build.stoagen.com/problems/07/index.md) - [08 Entity matching was degenerate](https://ethiopia-build.stoagen.com/problems/08/index.md) - [09 Wikidata has no Amharic labels for most concepts](https://ethiopia-build.stoagen.com/problems/09/index.md) - [10 The one ASR attempt died on a missing DLL](https://ethiopia-build.stoagen.com/problems/10/index.md) - [11 The chyrons have proper nouns spelled right](https://ethiopia-build.stoagen.com/problems/11/index.md) - [12 Singing and gemination](https://ethiopia-build.stoagen.com/problems/12/index.md) - [13 Script and encoding at the edges](https://ethiopia-build.stoagen.com/problems/13/index.md) - [14 Calendar and clock](https://ethiopia-build.stoagen.com/problems/14/index.md) - [15 Rights and provenance of anything translated](https://ethiopia-build.stoagen.com/problems/15/index.md) ## Optional - [Full corpus](https://ethiopia-build.stoagen.com/llms-full.txt): every page mirror in one file (alias full_site.txt), for tools that can take one large fetch. - [Feed](https://ethiopia-build.stoagen.com/feed.xml): pages by last update. - [Stoagen](https://stoagen.com/llms.txt): the system this site is published with.