Memory and its sources
PacoMind retains source records with the speaker and time. It can recall relevant memory before a reply and attach corrections to earlier information. Changing a search index preserves the original records.
01Open source research project
PacoMind stores an agent’s memory, identity and shared work across conversations and model changes. Built for Hermes, it supports local inference and is part of our ongoing, open source research toward PAGI (Pseudo-AGI).
MIT license · Phase 1 development and validation · Formerly ApsimoAGI
Request illustration
Hermes runs the agent, PacoMind stores memory and shared work, and selected models perform inference.
A message arrives through a configured conversation channel.
01 / Research lab
Ongoing, open source research toward PAGI (pseudo-AGI).
Messages arrive through Hermes channels. A person can provide context, correct the agent’s understanding and steer work across conversations.
Conversations → Hermes runtimeHermes handles conversations, model requests, tools, workers and schedules. PacoMind connects through plugins, a memory provider and request hooks.
Hermes ↔ PacoMind · Hermes → models and toolsHermes executes tools, workers and scheduled tasks. PacoMind keeps shared work state available across conversations, so users can inspect or redirect it within the deployment’s permissions.
Hermes → tools, workers and schedulesPacoMind retains original evidence and attributable corrections, recalls relevant context, and keeps identity and shared work across sessions. The deployment stores these records independently of the selected model.
Hermes ↔ memory and work context · PacoMind → selected modelsHermes and PacoMind each use their configured models. Local inference supports normal operation without a cloud model; changing a model does not replace the agent’s stored memory or identity.
Hermes + PacoMind → model inferenceIn development and validation. Read the architecture
Messages arrive through Hermes channels. A person can provide context, correct the agent’s understanding and steer work across conversations.
Conversations → Hermes runtimeHermes handles conversations, model requests, tools, workers and schedules. PacoMind connects through plugins, a memory provider and request hooks.
Hermes ↔ PacoMind · Hermes → models and toolsHermes executes tools, workers and scheduled tasks. PacoMind keeps shared work state available across conversations, so users can inspect or redirect it within the deployment’s permissions.
Hermes → tools, workers and schedulesPacoMind retains original evidence and attributable corrections, recalls relevant context, and keeps identity and shared work across sessions. The deployment stores these records independently of the selected model.
Hermes ↔ memory and work context · PacoMind → selected modelsHermes and PacoMind each use their configured models. Local inference supports normal operation without a cloud model; changing a model does not replace the agent’s stored memory or identity.
Hermes + PacoMind → model inferenceIn development and validation. Read the architecture
Project
PACO stands for Persistent Autonomous Cognitive Orchestration. Current work covers memory, agent identity and task coordination, with initiative and learning still under development.
PacoMind retains source records with the speaker and time. It can recall relevant memory before a reply and attach corrections to earlier information. Changing a search index preserves the original records.
A deployment can assign different models to conversation, reasoning and vision, including local models. Agent memory is stored separately, so it remains when a model changes.
PacoMind stores private identity, preferences and commitments. Hermes runs conversations, tools and workers. Connected sessions can inspect, redirect or stop shared tasks while other conversations continue.
Phase 1 is being developed and tested. Memory integration and shared task controls are implemented. Background initiative and autonomous improvement still need evidence from ordinary use. PAGI remains the research goal.
Contact