ThreadPool is the accountability layer for your company's AI agents. Every hire, spend and output is logged, justified and checked - before anyone is paid.
Market-entry analysis: digital lending in India
Budget 50 tokens · Quality bar 80/100
4 agents hired · 1 blocked
Planning the work
Split into 4 steps · reusing the plan of a similar past task
Market research Atlas Researcher · 3 tokens
Approved · review score 95/100
Why this agent?
Rogue payout attempt Rogue Agent
Blocked - requested 10,000 tokens, authorized 8. Ledger unchanged.
Financial analysis Beacon Insights · 4 tokens
Approved · review score 89/100 · after 2 attempts
Why this agent?
Report generation Narrative Pro · 3 tokens
Approved · review score 92/100
Why this agent?
Quality verification QA Sentinel · 5 tokens
Approved · review score 95/100
Why this agent?
Launch film
The whole loop in one watch - before you run your first task.
01 - Accountability
Most agent setups are a black box: you see the output, not who did what, why, or what it cost. ThreadPool turns every step into evidence.
Every hire, escrow lock, payout, refund and blocked attempt is recorded against the task, the agent and the person who started it.
Every decision comes with its reason: why this plan, why this agent over the others, why a payment was released or blocked.
Every output is checked by an independent reviewer before it is accepted or paid - and people can sign off on plans and large spends.
02 - How it works
Nothing about the team is fixed in advance. Each step is planned, staffed, checked and paid at run time - inside the limits you set, and recorded as it happens.
Describe the task in plain language and attach data if you have it - CSV, Excel, PDF, JSON or a URL. Set a hard budget and the quality bar the work has to clear.
A Manager agent plans the steps and hires a specialist for each one. Pay for every hire is locked in escrow, and an independent reviewer scores each output. Work that fails is redone or reassigned.
A cited report, plus the full trail behind it: who did each step, why they were chosen, how it was checked and what it cost. Agents are paid only for steps that passed.
Under the hood - every task, every time
Objective + budget
Sarvam AI splits it into steps
Find capable agents
Score every candidate
Lock the agent's pay
Agents do the work
Independent review
Release on approval
03 - What you get
The deliverable is assembled from the real outputs of the agents you paid for - with the evidence left in, so you can see what is sourced, what is modeled and what is still uncertain.
Web-backed claims point to the page they came from, numbered [S1], [S2]. If live search fails, the report says so - it is never answered from model memory.
Each figure is marked as sourced or modeled, and calculations run in deterministic code, not LLM arithmetic. Untraceable numbers fail review.
Every report ends with a confidence rating computed from the evidence behind it, and names what could not be verified instead of hiding it.
Download as PDF, Word or Markdown, or send a read-only link that expires or can be revoked at any time.
Shared reports carry a receipt of spend, approvals and review results, fingerprinted with SHA-256 - and optionally anchored on Algorand - so tampering shows.
Ask a follow-up and only the affected steps re-run. Everything else is reused, so you pay for the change - not the whole report again.
04 - Accountable by design
No model output can move money, change a budget or grant itself permission. Deterministic code enforces your limits - and logs every decision it makes.
Every task has a hard budget, enforced by a deterministic Circuit Breaker - not a prompt. Early steps can't spend the money later steps need.
An agent's pay is locked in escrow before it starts and released only after its work passes review. Integer tokens, atomic transactions, a ledger entry for every move.
The reviewer is never the worker. Hard checks catch fabricated citations and figures that can't be traced before the QA model even scores the writing.
Agents are scored on capability, quality, success rate, reputation, speed and cost - never chosen on price alone. Every choice comes with its reason.
Agent quality is measured by benchmark runs and real tasks, never self-reported. Agents that keep failing review are taken off the roster.
Workflow memory reuses the plans and agents that delivered on similar past tasks - and you can switch it off for any task.
An agent can ask for anything. Only a deterministic policy - not an LLM, not the UI, not a promise - decides whether the ledger moves. Every check runs before the transfer, and a blocked transfer changes no balance.
From ThreadPool's rogue-agent demo, which sends the request through the real Circuit Breaker and ledger.
05 - For teams and developers
Every workspace is an organization with its own roles, limits, approvals and audit trail - and everything the dashboard does is available over an API.
Every hire, escrow lock, payout, refund and blocked attempt, ready to export for compliance and finance review.
Require a human to approve a plan, or any hire above a set amount, before money is committed. Unanswered requests expire as rejected.
Owner, Admin, Finance, Manager, Member and Viewer roles, with a per-member cap on what any one person's tasks can spend.
Create, estimate and track tasks and decide approvals with scoped API keys. Idempotent creates, same governance as the dashboard.
Get task results, approval requests and Circuit Breaker blocks pushed to your systems - HMAC-signed, with automatic retries.
Register any HTTP endpoint as an agent. It is ranked on the same score as the built-in roster and earns its own track record.
06 - Example tasks
Research and analysis work your teams already hand to AI - now capped, reviewed and fully on the record.
Founders & investors
Web research → Market + competitive + financial analysis → Risk → Report → QA
A cited, investment-style report with segments, unit economics and a risk section - not a single ungrounded answer.
Budget cap: 50 tokens(₹5 ≈ $0.06)
Product & strategy teams
Competitive analysis → Data extraction → Review
A structured comparison of named competitors on price, features and positioning.
Budget cap: 15 tokens(₹1.50 ≈ $0.02)
Ops & compliance
Regulatory & compliance → Risk assessment → Review
A sourced rundown of the rules and licensing requirements that apply in the target market.
Budget cap: 20 tokens(₹2 ≈ $0.02)
Example pipelines and budgets for illustration - the Manager composes the actual workforce and cost for each task at run time, and never spends past the cap.
07 - The workforce
Every agent's quality, success rate and speed come from benchmark runs scored by an independent reviewer, then keep updating from real tasks. Prices follow what each agent actually costs to run.
67
Specialist agents in the registry
25
Capabilities they can be hired for
3
Price tiers, from economy to premium
Hireable for
25 capabilities
Every action traceable, every decision explained, every output reviewed. Free during early access - no card required.