Technical documentation, case studies, visual media, and funding context for the Starcaller platform.
Architecture overviews, security model, and Oracle reference.
Prism Nexus v2 — 8-Oracle orchestration with ICID constitutional governance. The OFP Kernel, Vault, and telemetry pipeline.
ArchitectureZero-trust architecture: memory provenance, cross-oracle signing, injection detection, physical action blacklist, Guardian subsystem.
SecurityComplete reference for all 8 Oracles: abilities, rings, invocation patterns, and inter-oracle communication protocols.
ReferenceCommand reference for the bot interface: /council, /morning, /task, /project, and slash-command integration.
Integration14 StarTeQ endpoints with authentication, rate limits, and example requests. teq.starcaller.uk
APIsThe six constitutional rules governing all Oracle behaviour: SAFETY, PRIVACY, FAIRNESS, TRANSPARENCY, ACCOUNTABILITY, HUMAN DIGNITY.
GovernanceStarcaller in daily operation — use cases that demonstrate the system’s breadth.
8 Oracles collaborate to produce a personalised daily briefing: weather, news, calendar, health, and security — synthesised by ICID in under 3 seconds.
Daily UseShi Gandang screens all IoT device communications. Physical action blacklist blocks dangerous commands at the kernel level regardless of which Oracle initiated them.
Home AutomationSia cross-references sources against citation networks while Nostradamus identifies trend trajectories. Cited, bias-audited research briefs delivered autonomously.
ResearchA simulated prompt injection via an IoT temperature sensor was blocked by the input sanitisation layer before it reached any Oracle. Audit record produced in 120ms.
SecurityEvery vault entry carries an HMAC-SHA256 source tag. An audit trace confirmed a recalled preference originated from direct user conversation, not a compromised device.
PrivacyYhi oracle, enhanced with financial capability, analyses spending patterns against energy-physics models. Physics-grounded recommendations, not generic advice.
FinanceStarcaller is applying for a £100k-£175k grant to fund the transition to true local inference on embedded hardware and to formally evaluate the privacy-preserving architecture.
For academic collaboration enquiries: sydney@starcaller.uk