01
Visible work
See what each coworker is doing now, what is queued next, which tools were used, and where a run got stuck.
Agent workspace for founders and operator-led teams
AI coworkers you can actually trust
Paperclip turns recurring work into specialized AI coworkers with visible queues, clear approval boundaries, auditable memory, and run history. Less “black box chatbot”. More dependable digital teammate.
Built for teams that want agents to do real work without losing visibility or control. Prefer a warmer founder-focused version? View agent factory for founders.
paperclip://agent-runs/live
Prototype ReviewerRUNNING
Sales Follow-up ScoutDONE
Content RepurposerQUEUED
Trust inspector
Roleprototype-reviewer
Can doread, critique, draft
Needs approvalpublish, email, spend
Modelclaude / codex / local
Memoryproject scoped, editable
Traceprompts, tools, outputs
Approval gate: this coworker can recommend changes, but cannot ship or contact customers until a human approves.
Why teams hesitate
Most agents ask for trust before they earn it.
They disappear into a long run, make tool calls you cannot inspect, remember things you cannot edit, and ask for broad permissions. Paperclip flips the model: every coworker is observable, scoped, and accountable by default.
02
Permission boundaries
Give each agent a role-specific sandbox: draft only, suggest, ask approval, or execute safe tasks inside a narrow scope.
03
Auditable memory
Project preferences, customer context, and decisions are visible and editable instead of buried in an invisible prompt.
How Paperclip works
Turn repeatable work into reliable coworkers.
Start with one recurring workflow. Paperclip helps define the role, tools, memory, approval gates, and success criteria — then gives you run history to improve it over time.
1
Describe the job
“Review every new prototype”, “draft sales follow-ups”, “summarize customer calls”, or any workflow you already repeat.
2
Set the trust contract
Choose what the coworker can read, write, spend, send, remember, and when it must ask for approval.
3
Watch the run
Each task gets a trace: prompts, tools, model, latency, outputs, decisions, and human approvals.
4
Promote what works
When a workflow proves useful, save it as a reusable coworker for your team or client workspace.
Coworkers, not chat tabs
Build a small AI team around real workflows.
Paperclip is for founders, growth teams, agencies, and technical operators who need agents that handle recurring work — not another blank prompt box.
PR
Prototype Reviewer
Reviews product prototypes, spots UX gaps, compares against strategy, and proposes the next experiment.
visible critiqueapproval before publishrun history
GT
Growth Teammate
Researches accounts, drafts outbound, repurposes product notes, and keeps follow-ups moving without spamming anyone.
draft onlyCRM contexthuman send gate
OP
Ops Analyst
Turns recurring dashboards, logs, meeting notes, and project updates into compact daily decisions.
scheduled runssource linkserror alerts
AI
Agent Builder
Helps power users package their own best prompts, tools, preferences, and guardrails as reusable agent templates.
templatesscoped memoryteam sharing
Trust is not a vibe. It is a product surface.
Paperclip makes agent work inspectable: what it knew, what it did, what it changed, what it wants permission to do next, and whether the outcome was useful.The trust layer
Everything important gets a trace.
Paperclip is designed for auditability from the start. Every coworker can produce a record of the model call, tool calls, memory used, approvals requested, files touched, and final output.
↳
Run timeline
A chronological view of agent reasoning milestones, tool usage, retries, queue events, and final delivery.
⌁
Model + cost visibility
Track provider, model, latency, tokens, errors, and cost estimates by coworker, workspace, or workflow.
⌘
Human-in-the-loop
Require approval for dangerous, expensive, external, or irreversible actions while letting safe work proceed.
Pricing
Start with one coworker. Scale when it earns trust.
Early pricing is designed for founders who want useful agents before they build a whole internal AI platform.
S
Starter
For one founder validating the first reliable AI coworker.
$49 / month
- 1 active AI coworker
- Agent launch checklist
- Basic approval boundaries
- Run history and weekly QA review
F
Founder Team
For small teams turning repeatable work into an agent factory.
$149 / month
- 5 active AI coworkers
- Reusable agent templates
- Approval gates for external actions
- Trace/audit timeline per run
- Daily and weekly routines
P
Partner
For agencies/operators managing AI coworkers across clients.
Custom
- Multiple workspaces
- Custom templates and QA scorecards
- Advanced audit and redaction policies
- Setup help for first workflows
Waitlist
Tell us what your first AI coworker should own.
We are inviting founders and operators with one painful recurring workflow: sales research, prototype review, support triage, content repurposing, or ops reporting.
Give your team AI coworkers with a paper trail.
Start with one workflow that already happens every week. Paperclip turns it into a scoped, observable, improvable coworker your team can rely on.