Available now
Persistent Researcher
Every source you research is kept, filed, and still there next week — not thrown away with the session that found it.
- harness
- hosted
- usage
- research
- source-filing
- citations
Agents
Agents That Do One Job Well
Specialized, proactive, autonomous. Each agent below runs a fixed workflow and returns evidence — pick the job, copy its prompt, and start.
The lineup
Pick the job, meet the agent
Every agent page carries its use cases, its when-to-use rules, and a prompt you can copy in one press.
Available now
Persistent Agent
Stops your coding agent from calling a job done at the first plausible answer.
- runtime
- local
- one-time
- persistence
- continuation-turns
- task-completion
Available now
Ensemble Agent
Gets several genuinely different approaches on the table and argued out before a single line is written.
- runtime
- local
- one-time
- ensemble
- multi-agent
- approach-exploration
Planned Task Board
Works an ordered list of tasks to the end, giving each one a fresh agent that has not been worn down by the last ten.
- runtime
- local
- one-time
- task-board
- ordered-tasks
- fresh-agent-per-task
Planned Stages Agent
Runs one job through phases that each get their own agent, their own instructions, and their own permissions.
- runtime
- local
- one-time
- stages
- phased-workflow
- per-stage-config
Planned Adversarial Team
Puts a team of hostile reviewers on your finished work before a real reviewer gets to it.
- runtime
- local
- one-time
- adversarial
- multi-agent
- red-team
How agents work
Three properties, every agent
Specialized
Each agent does one job with a fixed workflow and a fixed deliverable. You pick the job, not the model settings, and what comes back has the same shape every time.
Proactive
Agents keep working past the point a chat agent hands control back: threads hold their sources, sessions keep pushing on the task, boards work to the end of the list. The work compounds instead of restarting.
Autonomous
Give an agent its task once and it runs the loop inside the budget and permissions you approved. You review the result, not the intermediate steps.
Install
Two prompts per agent, not an install guide
Every agent page carries two prompts. The first one turns your job into the right invocation for that agent. The second one is setup instructions written for the coding agent you already use, so it installs the CLI, stores your key, and verifies the host while you only handle the three browser steps it cannot do for you.
1. Open the agent page 2. Copy its use prompt or its install prompt 3. Paste into your agent