Concepts
Architectural commitments, R&D, and the frameworks Hussh decisions run on.
89 pages
- One ADK — Agent Development Kit FrameworkThe architectural framework and runtime standard for building, orchestrating, and executing specialized personal agents under Hussh One.
- Intelligence Over RulesWhere a Hussh agent has been asked a question, a deterministic rule may inform the answer but may not replace it. Three forbidden shapes, and the two exceptions.
- Trusted Connections — Generalized Social GraphPublic concept overview of Hussh's privacy-first, directional social graph architecture for agentic trust and recipient-scoped interactions.
- World Model — user-side personal-knowledge storeNorth-star framing for user-owned personal context. Current repo truth is PKM/vault plus encrypted scoped exports; local/on-device claims remain future-state unless proven.
- Hermes Agent — the One runtime (Nous Research framework)Open-source agent framework from Nous Research that Hussh forks as the runtime for One. Provider-neutral, MCP-native, skills-driven — the concrete engine behind the BYOA commitment.
- Media CDN Offload PipelineArchitecture and design for offloading large public media assets to Cloud Storage and Cloud CDN.
- Partnership Claims StandardHow 🤫 talks about partners: platform independence and active partner pursuit are two different claims, and both must be stated accurately.
- Connected Systems, user-controlled integrations under PCHPThe architectural concept and design principles behind user-controlled integration surfaces (Connected Systems) brokered by the control plane and managed via the central One Dashboard, where the integrations live on your own data plane and the control plane holds none of what flows through them.
- Enterprise CRM RegistryUser-controlled enterprise data integration platform under PCHP governance, utilizing secure local-only mappings, schema-driven field interfaces, and zero-knowledge record storage.
- Personal Knowledge Model (PKM) Upgrade RehearsalPublic concept page defining the zero-information-loss upgrade and rehearsal framework for Hussh Personal Knowledge Models (PKM), framed on Hussh's control-plane posture.
- Source Library, Device-Bound Local Custody & Private PKM CapabilityPublic concept specification for the Source Library capability, defining device-bound local custody and private PKM capability bounds under PCHP, framed within the Hussh control-plane model.
- Gemini Runtime Configuration — AI Access & Model GovernanceHow Hussh One governs AI access and Gemini model execution: managed Vertex AI vs. BYOK (Google AI Studio / Google Cloud Vertex), PCHP consent boundaries, workload tiering, and privacy-preserving credential probing.
- 🤫 Private Agent One — mission and purposeThe purpose Private Agent One exists to serve: Americans connecting, communicating, and conducting commerce, charity and well-being work across their communities through human agents.
- Agents agreeing on a time — the most-asked use caseHow 🤫 Private Agent One settles a meeting time across several people by having agents negotiate over MCP, without anyone handing over their calendar.
- The Engineering BarHow 🤫 writes code: the Jeff Dean and Andrej Karpathy standard, and the production failures that taught us each rule.
- One Activity Feed — Cross-Domain Activity SurfaceThe One Activity Feed is the unified, cross-domain activity surface in Hussh One, providing a calm, paginated, read/unread-tracked stream of events across Consent, Location, Kai, KYC, Connected Systems, and Connections.
- One DocuSign and Fund Setup WorkflowVendor-neutral agreement execution and fund-setup workflow under One, with Nav/Connections authorization, trusted action confirmation, and e-signature provider options.
- Serialized Interaction RuntimeThe public design concept for Hussh's serialized interaction runtime: coordinates and sequences tap, voice, and native inputs across all client surfaces.
- AI & Search Engine Discoverability FrameworkPublic design and guidelines for Hussh One's discoverability architecture, aligning traditional SEO with Answer Engine Optimization (AEO) for LLMs and crawler-friendly discovery.
- Alpha Bets 27 — the triple filterFund A's universe filter. Three criteria — massive absolute FCF ($50B+), reinvestment runway at attractive returns, structural indispensability.
- DNSA Engine — entry / exit signalsFund A's signal layer. Non-proprietary discipline tools — RSI, VIX, put/call, credit spreads, IV rank, Bollinger position. The edge is execution discipline, not sophisticated signals.
- Hussh Discovery PipelineThe automated SEO, sitemap, and IndexNow discovery engine for the Hussh Search Console.
- Iron Rules — Fund A's five constitutional provisionsThe five non-negotiable risk and structural rules that govern Fund A — leverage ceiling, reserve floor, concentration cap, permanent HWM, GP co-invest.
- One Systems Meta HubThe technical HOW behind 🤫 Agent One — latency, failure modes, privacy boundaries, and scale.
- One Voice Onboarding Journey — first-run setupPublic concept page defining the stages, visual map, and privacy boundaries for One's interactive first-run voice onboarding journey.