Skip to main content

Introduction

Mockingbird is a time series controlled AI systems coordinator & workflow executor system, data indexer and searcher, that builds control loops and hierarchical planning rules around higher order goals & objectives using automated evaluation systems and model-to-model communication via synchronized quorums. It is designed to be a general purpose system that is eventually capable of becoming generally intelligent, and then super intelligent.

Usage

Mockingbird is an active research and development project, that is now in Beta. We're allowing free public usage of a curated set of features as we evaluate the system during this phase.

Costs

Mockingbird is currently free to use for self-service features while in beta. Though we may restrict account usage if we see abuse. If you connect your own OpenAI API key, you will be responsible for any costs associated with that, otherwise we will use our API key and then charge you 1-to-1 for the cost of the API usage that we incur, however you may have your usage throttled if you exceed our usage limits while using our API key.

Features Available via Self Service UI

  • Social Media Indexing (Twitter, Reddit, Discord)
  • Retrievals (API, Social Media)
  • Analysis/Aggregation Tasks
  • Integrated Time Window Search by Aggregation Windows
  • Multi-model communication and I/O
  • Time series controlled AI workflows, with dynamic time step control
  • Workflow + relationship builder (UI)
  • Integrated OpenAI + Social integration APIs, natively via Platform Secrets
  • Run Scheduler/Controller
  • Run Step-by-Step Cycle Execution History
  • Run Token Usage
  • Human-in-the-loop control trigger actions like API requests
  • Automated evaluation systems
  • Integrated JSON schema builder

Features Available To Researchers & Partners

  • Additional indexing sources (News, Blogs, Forums, etc)
  • Adaptive time series control via Metrics + Evals (PromQL, Adaptive)
  • Analysis cache feedback searching (eg. reviewing previous model generated analysis to steer actions)
  • Objective building for complex decision making, and milestone tracking
  • Decision tree generation for structured decision making
  • Unrestricted trigger actions (eg. outside of Human-in-the-loop control)
  • Experimental distributed executive task planner
    • Tasking single-to-multi-model task planning
    • Tasking scheduling, recursive workflows, objective overrides
    • Shared workflows contexts, selective contexts, searchable contexts
    • Scenario driven workflow generator
  • Multi-Model-Single-Model communication via quorums
  • Multi-Model-Multi-Model communication via quorums + task planning + synchronization
  • Ranked/Promotable-Model-Leader-Model-Follower
  • Multi-Model Team vs Multi-Model Team Competitive Scenarios
  • Multi-Modal (Text, Image, Video, Audio) analysis
  • AI driven compute infrastructure, building, purchasing, maintenance, & monitoring systems for Kubernetes + Cloud Vendor APIs
  • PagerDuty integration

Any questions or want access to research features? Contact us at support@zeus.fyi