The TOYTOCOL platform
TOYTOCOL is built around governed investigative work—not opaque answer generation. It keeps questions, evidence, reasoning, and human decisions connected so understanding can be inspected.
Core lifecycle
The stages are connected, but they are not automatic. Each one adds structure, context, or accountable review.
Something is noticed, reported, recorded, measured, or otherwise identified as potentially relevant.
An observation becomes something worth structured attention rather than remaining isolated information.
A bounded question creates a governed workspace for participants, evidence, relationships, reasoning, and findings.
Sources, records, and supporting material are attached to investigative context with provenance and relationships.
A bounded conclusion emerges from the evidence and reasoning available to the investigation.
Human review records whether a Finding is accepted within clearly stated limits. Verification is not absolute truth.
Verified conclusions can form the basis for reusable, traceable knowledge. This is the lifecycle destination—not a claim of an autonomous or persisted truth engine.
Conceptual destinationInvestigation workspace
TOYTOCOL brings investigative material together around a bounded Investigation. The surfaces reinforce one another instead of behaving like disconnected tools.
AvailableEvidence relationships
Evidence remains attached to investigative context, including source information, supporting material, activity, and associated findings. Explicit relationships make it possible to show how material supports, challenges, or contextualizes analysis without turning those links into automatic truth.
Knowledge graph
FoundationTOYTOCOL represents investigative entities as nodes and explicit relationships as edges. Investigation-scoped graph data can expose connections and paths without declaring that those connections prove truth.
Current boundary: the authorized Investigation-scoped graph backend exists, while the live graph renderer inside the Investigation Workspace is not yet connected.
Inspectable reasoning
TOYTOCOL keeps structural evaluation separate from persistent human analytical reasoning.
Deterministic stages inspect context validation, evidence readiness, relationship structure, finding coverage, and overall readiness. The result describes structure—not truth or confidence.
Persistent sessions let accountable participants record objectives, hypotheses, supporting or challenging arguments, evidence references, qualitative confidence, conclusions, and explicit limitations.
Verification and governance
Governance is expressed through public-safe principles rather than hidden behind a general claim of security.
Access follows governed resource scope and participation rather than interface visibility.
Confirmed and unconfirmed decisions remain distinct from reasoning confidence.
Current Finding state and recorded decision history remain inspectable.
Limitations and unresolved questions can remain visible alongside conclusions.
Public architecture boundary
TOYTOCOL follows an inward architectural direction: the web experience composes application capabilities, application work is constrained by domain rules, and infrastructure implements governed persistence and external boundaries.
This deliberately omits deployment topology, network detail, infrastructure identifiers, and operational configuration.
Explore the capability families
The Capabilities page separates current investigative functions from graph, extensibility, and ecosystem foundations.
View capabilities