Agents that learn from every conversation and get better over time.

Install in one command. They build their own knowledge pages, remember what you like, and keep themselves current — across Claude Code, Claude Chat, OpenClaw, and NemoClaw.

Create your own →Browse agents

npx @vectorize-io/self-driving-agents install marketing/seo --harness claude-code

Pick your harness above, or replace the agent slug with any template you want.

How it works

You never tell the agent to "save" or "remember." It decides what matters and keeps itself current.

You chat with the agent

Use it like any normal assistant. Ask questions, give feedback, share context.

Memory is retained

Conversations are auto-summarized into Hindsight as durable, structured memory.

Knowledge pages update themselves

After each session the agent rewrites the pages it owns — playbooks, references, preferences.

Next session it remembers

On the next chat the agent reads its current pages — it knows what works and what you prefer.

Want the deeper picture?

See how the harness, the agent, and Hindsight fit together — what crosses the boundary, what stays server-side, and why pages and memories live in two layers.

Available agents

Install the whole department or pick a specialty.

See all agents →

academic5 files

Academic

Extract scholarly arguments, source citations (primary vs secondary), methodological choices, period/regional context, contested claims, and the user's research questions and writing voice.

HumanitiesSocial Science

academic/humanities3 files

Humanities

Extract scholarly arguments, source citations (primary vs secondary), period and regional context, contested claims, and the user's research questions across history, anthropology, and narratology.

Research StyleOpen Questions

academic/social-science2 files

Social Science

Extract empirical findings, methodological choices, dataset limitations, and theoretical frames across psychology and human geography. Capture the user's research design and confidence in claims.

Method & EvidenceActive Questions

design8 files

Design

Extract design decisions, brand guidelines, visual systems, UX research findings, accessibility constraints, design critique outcomes, and stakeholder feedback on creative work.

UxVisual

design/ux3 files

Ux

Extract UX research findings, IA decisions, interaction patterns, usability test outcomes, and UI component standards. Capture stakeholder feedback and accepted vs. rejected proposals.

User InsightsInteraction Patterns

design/visual5 files

Visual

Extract brand guidelines, visual systems, image-direction prompts, accessibility/inclusive-visuals constraints, and storytelling decisions. Capture critique outcomes and locked vs. flexible elements.

Brand & Visual SystemCreative Direction

engineering29 files

Engineering

Extract architectural decisions, API contracts, performance numbers, incident root causes, code review patterns, deployment outcomes, and technical trade-offs across backend, frontend, infrastructu…

AiArchitectureBackendFrontendOpsSecurityWorkflow

engineering/ai5 files

Ai

Extract AI-engineering decisions: model choices, optimization patterns, voice/email integrations, and remediation outcomes.

Model & PromptAI Integrations

engineering/architecture3 files

Architecture

Extract software-architecture decisions, rapid-prototyping patterns, and senior-developer mentorship outcomes.

Architecture PillarsPrototyping Loop

Harnesses

Same agent, different runtimes. Pick the one you already use.

Compare harnesses →

Or roll your own agent.

An agent is just a directory: drop a bank-template.json next to some markdown files and point the CLI at it. Same self-driving behavior, your domain.

Read the spec →

text

my-agent/
  bank-template.json     # memory bank + knowledge pages config
  playbook.md            # any .md/.txt becomes seed knowledge
  advanced-tips.md

npx @vectorize-io/self-driving-agents install ./my-agent --harness claude-code