Roadmap - Atthene Documentation

Atthene Agents Studio is in closed beta. Your feedback directly shapes what we build next.

Your Input Shapes This

This roadmap shows our direction, but we prioritize based on what you need. If something here doesn’t work for your use case, or if we’re missing something critical—tell us.

Shape the Future

Have ideas? Email us at [email protected] with feature requests, timeline concerns, or integration needs.

Current Version: v0.2.1

Status: Live in closed beta | Released: November 2025

View Release Notes

See detailed release notes for v0.2.0 and v0.2.1

What’s Available Now

Core Configuration

YAML Configuration: Declarative agent definitions
Playground: Real-time testing with Agent Preview
Prompt Library: Versioned, reusable prompts
Built-in Tools: Pre-configured capabilities

Knowledge & Multi-Agent

Knowledge Base: File upload and document retrieval
Multi-Agent Workflows: Coordinated agent orchestration

Integration & Deployment

REST API & WebSocket: Integration endpoints with streaming
Public Chatbot Embedding: JavaScript widget for websites
Execution Tracing: Monitor agent behavior and performance

v0.2 is now live! Enhanced YAML configuration, conditional edges, structured outputs, UKnow integration, OCR/vision capabilities, and expanded model provider support are all available.

Experiencing Issues?

Report bugs at [email protected] — we’re actively fixing issues based on alpha feedback.

v0.2: Extensions & Intelligence

Status: Released | v0.2.0: November 21, 2025 | v0.2.1: November 25, 2025

View Release Notes

See what’s new in v0.2.0 and v0.2.1

Model Provider Support

Choose your LLM provider per agent:

Mistral

Qwen

More...

Need a specific model provider? Let us know which ones matter for your use case.

Advanced YAML Features

Config Reference
Input/Output System
Conditional Edges
Structured Output

Use variables and reference configs across agents:

variables:
  api_endpoint: "https://api.example.com"
  
agents:
  - name: "agent1"
    config_ref: "{{ variables.api_endpoint }}"

Pass data between agents explicitly:

agents:
  - name: "analyzer"
    outputs: ["analysis_result"]
  
  - name: "responder"
    inputs: ["analysis_result"]

Dynamic routing based on agent outputs:

edges:
  - from: "classifier"
    to: "urgent_handler"
    condition: "output.priority == 'high'"
  
  - from: "classifier"
    to: "standard_handler"
    condition: "output.priority == 'normal'"

Enforce JSON schemas for agent responses:

agents:
  - name: "data_extractor"
    structured_output:
      type: "json_schema"
      schema:
        type: "object"
        properties:
          name: {type: "string"}
          age: {type: "number"}

Live Avatar Chatbots

New embedded chatbot format with talking avatars:

Visual Features

Default avatar library included
Custom avatars available
Real-time lip-sync
Dynamic expressions

Audio Features

Speech-to-text input
Text-to-speech responses
Natural conversation flow
Multiple voice options

Perfect for customer-facing applications where visual presence matters.

Want Custom Avatars?

Reach out to discuss your branding needs.

Model Context Protocol (MCP)

Integration with the MCP standard for:

Standardized Tool Definitions

Use industry-standard tool formats

Cross-Platform Compatibility

Work seamlessly across different AI platforms

Community Tool Sharing

Access and share tools with the community

Easier Custom Tool Creation

Build tools faster with standard templates

Cloud Knowledge Base Connectors

Connect cloud storage as knowledge sources:

Google Drive

Dropbox

OneDrive

SharePoint

AWS S3

Azure Blob

Real-time sync and permission-aware retrieval included.

Which Cloud Provider Matters Most?

We’re prioritizing based on demand. Email us your top 3 cloud integrations.

Tool Library Integration

Comprehensive tool ecosystem for agent capabilities:

Pre-built Tools

Communication platforms
Project management systems
Document repositories
Cloud storage services
CRM and sales tools
Development platforms

Integration Features

OAuth authentication flows
Full CRUD operations
Real-time data sync
Permission-aware access
Rate limiting compliance
Error handling & retries

Agents can seamlessly interact with your existing tool stack through standardized integrations.

What Tools Do You Need?

Tell us which tools are essential for your workflow at [email protected]

v0.3: Production Readiness

Timeline: Dec 2025

Focus: Evaluation, control, advanced orchestration

Sophisticated Orchestration Examples

Pre-built templates for advanced patterns:

Map-Reduce Workflows

Distribute work across agents and aggregate results

Parallel Processing

Run multiple operations simultaneously with intelligent aggregation

Multi-Stage Research

Iterative information gathering and analysis pipelines

Approval Chains

Sequential decision-making with fallback mechanisms

Recursive Loops

Self-improving agent processes

Deep Research Agent

Specialized agent template for comprehensive research tasks:

Multi-Source Gathering

Pull information from diverse sources automatically

Iterative Refinement

Continuously improve query quality based on results

Source Verification

Validate information and maintain citation trails

Report Generation

Create structured findings with actionable insights

Evaluation Framework

Quality Scoring

Automated agent response quality assessment

A/B Testing

Compare prompt variations for optimization

Benchmarking

Track performance metrics over time

Cost Tracking

Monitor spending per agent and conversation

Human-in-the-Loop (HITL)

Enable human oversight for critical decisions:

Approval Workflows

Require human approval for sensitive actions before execution

Interactive Decision Points

Pause workflows for user input at key moments

Feedback Collection

Gather user insights during agent execution

Manual Overrides

Take control and redirect agents when needed

Parallel Agent Workflows

Run multiple agents simultaneously:

agents:
  - name: "researcher_1"
  - name: "researcher_2"
  - name: "researcher_3"
  - name: "aggregator"

edges:
  - from: "__start__"
    to: ["researcher_1", "researcher_2", "researcher_3"]
    mode: "parallel"
  
  - from: ["researcher_1", "researcher_2", "researcher_3"]
    to: "aggregator"
    mode: "collect"

Dramatically reduce total workflow execution time

Context Engineering

Intelligent context management:

Automatic Optimization

Smart context window management without manual intervention

Conversation Summarization

Maintain context while reducing token usage

Priority-Based Inclusion

Keep the most relevant information in context

Token Usage Optimization

Minimize costs while maximizing effectiveness

v0.3 feature priorities will shift based on what v0.2 alpha users need most. Stay vocal about what matters.

v0.4: Autonomy & Scale

Timeline: Jan 2026

Focus: Goal-oriented agents and automation

Planning Agent

New agent type with advanced capabilities:

Goal Decomposition

Break complex objectives into actionable steps

Dynamic Planning

Create execution plans on the fly

Adaptive Execution

Adjust plans based on outcomes

Sub-Agent Coordination

Orchestrate multiple agents toward goals

Goal-Based Workflows

Agents work toward defined objectives with incentive structures:

agents:
  - name: "sales_assistant"
    goals:
      - objective: "schedule_meeting"
        priority: "high"
      - objective: "gather_requirements"
        priority: "medium"

Intelligent Agent Collaboration

Advanced coordination methods:

Negotiation
Consensus Building
Resource Sharing
Conflict Resolution

Agents discuss and reach agreements on approach

CronJobs & Trigger Events

Automation capabilities:

Schedule-Based

Execute agents on defined schedules

Webhook Triggers

React to external system events

Event-Driven

Respond to emails, tickets, and more

Monitoring

Automated alerts and notifications

Beyond v0.4

Agent Marketplace

Community-shared agents and workflows

Self-Improving Agents

Learn from past interactions and optimize over time

Multimodal Support

Process images, audio, and video

On-Premise Deployment

Self-hosted for maximum control and security

Enterprise SSO

SAML/OAuth integration for enterprise authentication

These are ideas, not commitments. Features depend on how v0.2 and v0.3 land with alpha users.

How to Influence the Roadmap

Use the Platform

The more you build, the more we learn what’s needed

Report What's Blocking You

If a missing feature stops your project, tell us immediately

Describe Your Use Case

Context helps us prioritize. What are you building? Who’s it for?

Vote on Features

When we ask “which integration matters most?”, respond. We listen.

Feedback Channels

Feature Requests

Email: [email protected]
Subject: “Feature Request: [Your Idea]”

Bug Reports

Email: [email protected]
Subject: “Bug: [What’s broken]”

Integration Needs

Email: [email protected]
Subject: “Need Integration: [Tool/Service]”

In-Platform Feedback

Click “Give Feedback” in the top right of the platform

Transparency Commitment

We’ll update this roadmap monthly as priorities shift. If something gets delayed or dropped, we’ll explain why.

Last Updated

November 2025

Next Update

December 2025 (post v0.3 release)

Alpha Disclaimer: Timelines are estimates. Features may change or be postponed based on technical complexity and user feedback. We prioritize building the right thing over hitting exact dates.

Have 5 Minutes?

Quick Survey: What’s the #1 feature that would make AMAS indispensable for your project?→ [email protected]

Getting started

Agent Configuration

Multi Agent Systems

Atthene Agents Studio

Knowledge Base

Platform

Integration

​Your Input Shapes This

Shape the Future

​ Current Version: v0.2.1

View Release Notes

​ What’s Available Now

Experiencing Issues?

​ v0.2: Extensions & Intelligence

View Release Notes

​ Model Provider Support

Mistral

Qwen

More...

​ Advanced YAML Features

​ Live Avatar Chatbots

Visual Features

Audio Features

Want Custom Avatars?

​ Model Context Protocol (MCP)

​ Cloud Knowledge Base Connectors

Google Drive

Dropbox

OneDrive

SharePoint

AWS S3

Azure Blob

Which Cloud Provider Matters Most?

​ Tool Library Integration

Pre-built Tools

Integration Features

What Tools Do You Need?

​ v0.3: Production Readiness

​ Sophisticated Orchestration Examples

​ Deep Research Agent

​ Evaluation Framework

Quality Scoring

A/B Testing

Benchmarking

Cost Tracking

​ Human-in-the-Loop (HITL)

​ Parallel Agent Workflows

​ Context Engineering

​ v0.4: Autonomy & Scale

​ Planning Agent

Goal Decomposition

Dynamic Planning

Adaptive Execution

Sub-Agent Coordination

​ Goal-Based Workflows

​ Intelligent Agent Collaboration

​ CronJobs & Trigger Events

Schedule-Based

Webhook Triggers

Event-Driven

Monitoring

​ Beyond v0.4

​ How to Influence the Roadmap

​ Feedback Channels

Feature Requests

Bug Reports

Integration Needs

In-Platform Feedback

​ Transparency Commitment

Last Updated

Next Update

Have 5 Minutes?

Your Input Shapes This

Current Version: v0.2.1

What’s Available Now

v0.2: Extensions & Intelligence

Model Provider Support

Advanced YAML Features

Live Avatar Chatbots

Model Context Protocol (MCP)

Cloud Knowledge Base Connectors

Tool Library Integration

v0.3: Production Readiness

Sophisticated Orchestration Examples

Deep Research Agent

Evaluation Framework

Human-in-the-Loop (HITL)

Parallel Agent Workflows

Context Engineering

v0.4: Autonomy & Scale

Planning Agent

Goal-Based Workflows

Intelligent Agent Collaboration

CronJobs & Trigger Events

Beyond v0.4

How to Influence the Roadmap

Feedback Channels

Transparency Commitment