Build mediation platforms, political analysis tools, negotiation assistants — powered by the TACITUS ontology. Don't let your users reason from blind prose. Give them structured conflict context — and watch them make better decisions.
Your LLM doesn't know what changed after a leadership transition. It can't distinguish a broken promise from a new commitment. It collapses months of negotiation into a flat context window. Plug into TACITUS and your app gains structured conflict memory — temporal ordering, causal chains, actor networks, source provenance. The structure that prose hides, served as an API.
We run it, you query
TACITUS handles the full lifecycle — ingestion, extraction, graph management. You access structured conflict intelligence via API.
SDKs, build features
Use our Python and TypeScript SDKs to build conflict-aware features directly into your application. Full access to ontology types and graph queries.
Self-host, MIT
Deploy the TACITUS pipeline on your own infrastructure. MIT licensed. Fork it, extend it, contribute back.
Three lines to structured conflict intelligence.
from tacitus import Client
client = Client(api_key="your-key")
graph = client.analyze("Two departments are disputing budget allocation...")
# Returns structured conflict graph
print(graph.actors) # [Actor(name="Engineering"), Actor(name="Marketing")]
print(graph.claims) # [Claim(text="Engineering needs 60% of Q3 budget", actor="Engineering")]
print(graph.interests) # [Interest(text="Ship v2.0 before competitor launch", actor="Engineering")]
print(graph.constraints) # [Constraint(text="Total budget is fixed at $2M")]import { Tacitus } from '@tacitus/sdk';
const client = new Tacitus({ apiKey: process.env.TACITUS_API_KEY });
const graph = await client.analyze({
text: "Two departments are disputing budget allocation...",
ontology: "conflict-v2", // 8 primitives, 41+ classes
});
// Typed response with full ontology
console.log(graph.actors); // Actor[]
console.log(graph.claims); // Claim[]
console.log(graph.edges); // TypedEdge[] with causal + temporal ordering// Ingest a document
const result = await tacitus.ingest({
source: 'complaint.pdf',
extractionMode: 'full-ontology'
});
// Query the graph
const actors = await tacitus.graph.query({
type: 'Actor',
conflictId: 'case-003',
include: ['claims', 'interests', 'commitments']
});
// Generate intelligence
const briefing = await tacitus.generate({
template: 'investigation-summary',
conflictId: 'case-003',
format: 'compliance-ready'
});Python + TypeScript SDKs — In development. Request early API access.