01Python · standard library
Runnable reference
Policy knowledge pipeline ↗
Exact sources → semantic layer → bounded capsule → cited draft. Includes an evidence-inspection agent and explicit cloud/local provider adapters.
Tacitus / AI data infrastructure
Policy expertise.
Structured for AI.
Deterministic data structures for policy and political analysis. Define the concepts, connect the evidence and give AI the context that matters.
Concept register01—18
Mandate≠ParticipationCommitment≠ProposalApplicability≠Adoption
Source wording. Stated conditions. Reviewable meaning.
Make meaning explicit.
Concepts, ontologies, relationships and weighted source records.
Explore the knowledge ↗02 / CapsulesBring the right context.
Task-specific sources, distinctions and expert checks for AI.
Inspect a capsule ↗03 / ArtifactsProduce useful work.
A briefing, comparison, meeting agenda or review note.
Open an artifact ↗The public code
Open the implementation.
02TypeScript
Local reference
Context workbench ↗
Inspect actual model drafts beside exact evidence, qualifications, graph objects and recorded agent steps.
03Rust
Experimental core
DIALECTICA ↗
Explore capsule contracts, typed context and ontology-backed preparation.
Start with the policy pipeline. Explore the other projects as local references and research code. Hosted PRAXIS integration and held-out quality evaluation remain in development. Inspect a real model run ↗
Why structure the knowledge?
Keep the distinctions
that change a briefing.
Policy work depends on attribution, changing identities and conditional language. Capture them once as source-linked records, then reuse them across questions and drafts.
See a diplomatic drafting example ↗- Identity
- Which actor is meant? An alias or affiliation does not establish shared authority.
- Evidence
- Which report is original? Repetition does not become independent confirmation.
- Language
- Which reservation matters? Keep “subject to instructions” attached to the statement.
- Reuse
- Which source and version support the next brief? Retrieve them together.
Start with the work
Where to start with
Policy / Political AI.
A recurring task. A clear output.
Evidence a reviewer can check.
If your team is facing…Start with…
Opening the task guide…
Explore the knowledge
behind a question.
Choose a use case. Inspect the published vocabulary, source passages and task-specific connections.
Opening the knowledge database…
Why this layer?
Open a layer to inspect its purpose.
Published vocabularySource materialTask & assessment
Explore nodes as a list
Saliency
Keep the evidence
that changes the answer.
Preserve dates, exceptions, attribution and missing inputs. Select complete evidence before supplementary instructions.
Efficiency
Spend context
on the actual task.
Reuse defined concepts and deduplicate sources. Measure context cost and output quality together.
Research, tests and limits ↗Deterministic means the same records and rules give the same structured result. It does not make political interpretation or model generation deterministic.
Questions worth asking
Why this work matters.
Why prepare policy knowledge before asking AI?
A useful policy answer depends on who said something, which actor they meant, the source version and the conditions attached. A knowledge pipeline captures those relationships once, so each briefing can reuse and inspect them. The intended benefit is less repeated preparation and fewer lost distinctions. Each workflow still needs evaluation.
How can data structures improve saliency?
Saliency means keeping the evidence that changes the conclusion. A task-specific structure can retrieve the relevant actor, clause and exception together, instead of sending an undifferentiated document pile. Models can still miss or misuse that evidence. Lost in the Middle shows that evidence position matters in the tested long-context settings; it does not validate Tacitus.
Where can efficiency come from?
Reusable definitions and source records reduce repeated assembly. Deduplication can avoid spending context on the same evidence twice. Selection can keep unrelated material outside the task. Building and maintaining a graph also costs time and compute, so we measure preparation, retrieval, model usage and review together. Shorter context alone is not success.
How can a graph preserve diplomatic language?
Keep the exact wording as an object linked to its speaker, source version, speech act, conditions and review state. Prepared to discuss, subject to instructions must remain distinguishable from agrees to establish. Our authored example shows how those records can guide a draft and expose a lost reservation. A graph does not prevent semantic averaging on its own; extraction, interpretation and generation need separate checks.
What has Tacitus actually benchmarked?
We have reproducible checks for evidence retention, context size, source identity and rules. A frozen 24 KB compiler regression now keeps three exact quotations and an artifact that the old version omitted. An earlier 48-response model evaluation had mixed results. We have not established a general improvement in generation quality, inference cost or analyst time.