01
Four knobs, one unit
Instructions, knowledge, tools, and model define how an agent behaves. Change one knob without re-training anything.
Agent conceptsAgents
What are agents in Tale?
A Tale agent is the four-knob combination of instructions, knowledge, tools, and a model. Editors build them once; the team runs them in chat, automations, and projects — on harnesses like Claude Code, Codex, Hermes, and OpenClaw when the work needs a coding CLI in a sandbox.
How agents show up in Tale
Dock the agents you trust, ground them in knowledge, run them in workflows, and scope them by project.
01 Projects
Scope agents by team
Each project keeps its agents, members, and knowledge in one workspace — support never sees engineering’s stack.
02 Knowledge
Ground every agent in what you know
The Support Agent gets macros, brand voice, and the help-center crawl — and nothing else. Each agent searches only the slices you grant.
03 Automations
Run agents in multi-step work
Workflows call agents as LLM steps with conditions and actions — every run leaves an execution log.
04 Sandbox
Watch the agent in Files and Live
Agents that run on a harness work in a sandbox — browse the workspace tree, open the file they just wrote, and watch the live browser as they click through the result.
What can agents do?
What one agent definition gives the whole team — and what it adds beyond a bare model.
02
Harnesses dock in
Bind Claude Code, Codex, Hermes, and OpenClaw as the harness a project agent or automation agent node runs on.
Harnesses03
Delegation between agents
Specialist agents hand work to each other when a thread leaves one domain — without losing the transcript.
04
Skills as reusable bundles
Package a writing voice or multi-step pattern once and bind it to up to ten agents.
05
Knowledge scoped per agent
Each agent searches only the library slices you grant — team docs stay invisible outside the team.
Agent knowledge
Agents FAQ
A Tale agent combines instructions, knowledge, tools, and a model into one reusable unit that Members run in chat and automations.
Yes. Claude Code, Codex, Hermes, and OpenClaw ship as harnesses — bind one when you create a project agent or an automation agent node.
Use an agent when the conversation stays in one domain and voice. Reach for an automation when the work needs multiple steps, branches, schedules, or approvals.
Yes. Every agent binds its own primary model and fallbacks — Claude Code can run on Claude Opus while a support specialist runs Claude Sonnet. Swap the model without rebuilding the agent.
See Tale on your stack
Book a guided demo or talk to the team about self-hosted deployment, pricing, and hardware.