Skip to main content

Dashboard Guide

The HatiData dashboard at app.hatidata.com is organized into four sections: Overview, Intelligence, Data, and Admin. This guide walks through every page.

SectionPages
OverviewCommand Center
IntelligenceAgent Fleet, Memory Store, CoT Replay, Branches, Triggers, Agent Keys
DataSQL Editor, Table Editor, Schema Browser, Visualizations
AdminTeam, Settings, Billing & Usage, Audit Log

Command Center (/dashboard)

The first page every user sees. It provides a real-time snapshot of your HatiData environment.

  • 4-stat grid at the top: active agents, memory records, CoT steps, and SQL queries today.
  • Running agents panel with direct links to each agent's CoT replay.
  • Live activity feed showing recent queries, memory writes, and trigger firings.
  • Project pipeline with phase badges and progress bars for tracking onboarding milestones.

A personalized greeting addresses the logged-in user by name.

Agent Fleet (/agents)

A live registry of every agent that has self-registered with HatiData.

Trust Levels

Each agent carries a trust badge:

LevelBadgeMeaning
ProvisionalGrayNewly registered, awaiting review
VerifiedBlueIdentity confirmed, standard permissions
TrustedGreenElevated permissions, can write to shared tables
OrchestratorVioletCan delegate tasks to other agents

Provisional Review Queue

When agents are pending review, a gold banner appears at the top of the page. Use the Promote or Restrict buttons to act on each agent.

Fleet Table

Columns: agent name + fingerprint, trust badge, state badge (active/idle/offline), framework, memory key count, and last active timestamp. Click any row to open that agent's detail view.

Memory Store (/agent-memory)

Browse and search all agent memories across your organization.

Memory Table

Each row represents a stored memory. Visual indicators help you triage at a glance:

  • Gold dot — pinned to hot storage (fast retrieval, higher cost)
  • Violet dot — cross-project global memory (shared across projects)
  • Green pulse — live write in progress

Use the Global search toggle to search memories across all projects, not just the current one. The search bar supports both keyword and semantic queries.

Chain-of-Thought Replay (/cot-replay)

Replay any agent's reasoning chain with cryptographic verification.

Timeline View

The page uses a two-column layout: the session list on the left and the trace detail on the right.

Each step in the trace shows:

  • Step type (e.g., reasoning, tool_call, observation)
  • Content summary
  • SHA-256 hash chip linking to the previous step
  • Verification status icon

A green Chain verified badge at the top of the trace confirms the entire chain is tamper-evident. If any hash is broken, a red warning appears instead.

Branches (/branches)

Copy-on-write branches let agents explore alternative paths without affecting production data.

Branch cards display: the owning agent, project, commit count, a progress bar showing materialized vs. view-only tables, and the branch state (active, merged, discarded). Click a card to inspect the branch's diff against main.

Semantic Triggers (/triggers)

Event-driven rules that fire when agent state changes match a semantic condition.

A gold Gate warning banner appears when any trigger requires human approval before proceeding. Each trigger row includes a toggle switch to enable or disable it without deleting the rule. Click a trigger to view its evaluation history and fire count.

Agent Keys (/agent-keys)

HatiData uses a two-tier key model:

  1. Org-level key — agents present this key to self-register with the platform.
  2. Agent-scoped keys — after registration, each agent receives a scoped key bound to its identity for tighter access control.

A concept card at the top of the page explains the model visually. Below it, a connection snippet shows the exact psql or SDK connection string agents should use.

SQL Editor (/sql-editor)

A Monaco-based SQL editor purpose-built for HatiData.

Features:

  • Syntax highlighting for HatiData SQL, including semantic_match() and semantic_rank() functions.
  • Run, Run Selection, and Explain buttons in the toolbar.
  • Results table with sortable columns and a Visualize tab to chart results inline.
  • Query history sidebar showing recent queries with execution time and row count.

Table Editor (/table-editor)

A spreadsheet-like interface for viewing and editing table data directly. Supports inline cell editing, row insertion, and row deletion. Changes are committed through the standard query pipeline, so all access control and audit logging still apply.

Schema Browser (/schema-browser)

A tree navigator organized as Database > Schema > Table. Expanding a table reveals its columns with data types, nullability, and any applied column masks. Use the search bar to filter across all schemas.

Visualizations (/visualizations)

Build charts from query results. Supported chart types include bar, line, area, scatter, and pie. Completed charts can be pinned to the Command Center for at-a-glance monitoring.

Team (/team)

Manage users in your organization: invite new members, change roles, or remove access.

RoleCapabilities
OwnerFull access, can delete organization
AdminManage users, settings, billing
DeveloperQuery, write data, manage agents
AnalystQuery data (read-only)
AuditorRead-only access to audit logs and CoT replays

Settings (/settings)

Configuration is organized into sub-tabs: General, Notifications, Security, Environments, and Danger Zone.

Agent Registration Policy (General tab)

Three options control how new agents join your organization:

PolicyBehavior
OpenAny agent with the org key is auto-promoted to Verified
Governed (recommended)Agents start as Provisional; admins review and promote
StrictAgents must be pre-registered by an admin before connecting

An auto-promotion threshold slider lets you set how many successful queries an agent needs before automatic promotion from Provisional to Verified (Governed mode only).

Billing & Usage (/billing)

View your current plan, usage metrics (queries, storage, agents), and invoices. Set budget alerts to receive notifications when usage approaches a threshold. Plan cards show feature comparisons across Free, Pro, and Enterprise tiers.

Audit Log (/audit-log)

A unified audit view that merges query audit events and admin audit events into a single timeline.

Filter by agent, user, event type, or time range. Each entry shows the actor, action, target resource, and timestamp. Use the Export button to download filtered results as CSV or JSON.

Stay in the loop

Product updates, engineering deep-dives, and agent-native insights. No spam.