Runtime Product
Freelee is the public UK Bot Marketplace runtime. Users register, define their scope, open a cockpit dashboard and work with published category instructors through narrative chat.
Documentation
Freelee UK Bot Marketplace
A public AI marketplace runtime for UK category instructors, connected to a private T1000 owner control plane for brand, bot and publishing operations.
T1000 Control Plane
Owner operations live in T1000. Branding, footer links, category marketing, master instructor bots, publishing and test chat are managed from t1000.website/admin/freelee.
Data Model
The runtime database stores users, sessions, preferences, categories, sectors, published agents and chat history. The T1000 master database stores draft instructor bots and prompt versions.
Access Boundary
Public users never access the owner tools. T1000 talks to Freelee through private service-token APIs, while Freelee validates and writes its own runtime records.
Current Build
Categories
20
UK marketplace categories
Sectors
103
Mapped category sectors
Runtime
Live
Login, dashboard and chat
Database
Freelee Control data ownership
Freelee uses a two-database architecture. The public runtime owns user-facing records in aigencydb. T1000 owns the private planning and publishing inventory in freelee_control. The browser never writes directly to either database; all writes go through server-side APIs.
Used directly by freelee.cv
Runtime database: aigencydb
organizations
Keeps the primary `ukbots` organization and org-level settings.
users
Stores registered users, owner/admin identity and password hashes.
memberships
Connects users to `ukbots` with owner/admin/member roles.
auth_sessions
Backs the secure HttpOnly `freelee_session` cookie.
user_preferences
Stores onboarding scope: categories, sectors, audience type, scenario and goals.
categories
The 20 UK marketplace categories imported from the source dataset.
sectors
The 103 category sectors used for future bot matching and specialization.
category_marketing
T1000-editable card copy, selling points, badges and visibility.
organization_site_settings
T1000-editable brand, logo, hero, footer, CTA and legal JSON.
organization_ai_settings
Owner technical defaults: audience mapping, fast/smart/deep configs and model choices.
agents
Published runtime bots that can appear in dashboard and `/chat/[agentSlug]`.
agent_categories / agent_sectors
Runtime taxonomy links connecting published bots to marketplace scope.
threads / messages
Persistent chat sessions and message history for user conversations.
Used by the owner control plane
T1000 master database: freelee_control
bot_profiles
Master bot records in T1000: slug, category, audience, status, card/persona/model JSON and publish state.
bot_prompts
Versioned current and historical system prompts for master bots.
bot_category_links
Master-side category metadata used before publishing into runtime.
bot_sector_links
Reserved for future sector-level bots beneath each category.
bot_publish_log
Records successful and failed publish attempts from T1000 to Freelee runtime.
Publish flow
Step 1
Owner edits brand, category cards, instructors, technical settings or runtime bots in T1000 Freelee Control.
Step 2
T1000 stores master instructor work in `freelee_control` and calls Freelee internal APIs with a service token.
Step 3
Freelee validates the payload and writes runtime-safe records into `aigencydb`.
Step 4
Published agents become visible to users through dashboard scope rules and the narrative chat route.