Many agents, one goal
A run hands parts of the work to subagents that work at the same time.
NewPer-run spend limits
Run your own AI agent
factory for
.
Agent swarms accelerate your workflow 24/7, around the clock. They adapt as your needs change and stay within your budget.
1
Factories/security
Waiting for you- asked just nowDay 14 of 141Safe probes sent
1,284
read-only, inside your scope
Confirmed findings
3
1 high severity
Routes checked
41
new or changed, 38 clean
Spend
$4.00
of $25.00 cost budget
Probes by checklast 14 days
1,284probes sent3 confirmed
Missing auth312Exposed data486Verbose errors174Security headers208Open redirects104
Agent tasksLatest first
| Task | Status |
|---|---|
| Find the latest deployment | Finished |
| Probe new API routes for exposed data | Working |
| Review 3 draft patch pull requests | Waiting for you- asked just now |
| Redact tokens from the evidence | Finished |
| Write a regression test for the auth fix | Working |
Factories/engineering
WorkingDay 14 of 14Draft PRs opened
62
each with a before and after test
Bugs reproduced
93%
the rest reported as unconfirmed
Existing issues
14
commented instead of duplicated
Spend
$8.68
of $25.00 cost budget
Draft pull requests3 of 4 opened as drafts
- #482Fix 500 on empty cart checkoutReproduce FinishedFix and test FinishedDraft PR Finished
- #485Handle expired session tokensReproduce FinishedFix and test FinishedDraft PR Finished
- #487Fix off-by-one in invoice datesReproduce FinishedFix and test FinishedDraft PR Finished
- #490Guard missing user in settingsReproduce FinishedFix and test WorkingDraft PR Idle
Agent tasksLatest first
| Task | Status |
|---|---|
| Read Vercel deployment events | Finished |
| Reproduce a checkout error | Working |
| Review 4 draft pull requests | Working |
| Write a failing test for the session bug | Working |
| Comment the fix on a Linear issue | Finished |
Run plan
Each factory run reads deployment events, open issues, and code, then opens a tested draft pull request for each bug nobody filed.
- Read Vercel events and Linear issues
- Reproduce each bug before fixing it
- Open each fix as a draft pull request
- Cost budget $25.00, warn at 70%
Ready for your review
Review 4 draft pull requests
Each one names the bug, reproduction steps, root cause, fix, and test results.
Factories/finance
WorkingDay 14 of 14Passages compared
12,479
against the last digest
Material changes
17
met your materiality bar
Digests filed
9
dated pages in Notion
Spend
$3.46
of $25.00 cost budget
Source readslast 14 days
1,023pages read14 material changes
Agent tasksLatest first
| Task | Status |
|---|---|
| Read central bank news | Finished |
| Compare sources with the last digest | Working |
| Review today's digest of 3 changes | Working |
| Estimate revenue at risk from Stripe | Working |
| Quote old and new text for each change | Finished |
Run plan
Each factory run watches market, supplier, and regulation sources, sizes the revenue at risk from Stripe, then files a dated, cited digest in Notion.
- Read your market, supplier, and regulation sources
- Keep only changes above your materiality bar
- Never move money or post entries
- Cost budget $25.00, warn at 70%
Ready for your review
Review today's digest of 3 changes
Each change quotes the old and new text, links its source, and states its revenue estimate.
Factories/product
WorkingDay 14 of 14Flow checks
412
read-only requests only
Failures found
23
traced to file and line
Replays matched
318
linked replays and error reports
Spend
$5.62
of $25.00 cost budget
Flow checkslast 14 days
405flow checks23 failures
- Sign up
- Checkout
- Search
- Settings
- Billing
- Passed
- Failed
- Not walked
Agent tasksLatest first
| Task | Status |
|---|---|
| Walk sign-up with read requests | Finished |
| Trace a checkout failure to its code | Working |
| Review 5 fix pull requests | Working |
| Write the end-to-end test for search | Working |
| Skip billing steps that write data | Finished |
Run plan
Each factory run walks your critical flows with read-only requests, then opens a pull request with the fix and an end-to-end test for each break.
- Walk critical flows with read requests only
- Trace each failure to file and line
- Open a pull request per root cause
- Cost budget $25.00, warn at 70%
Ready for your review
Review 5 fix pull requests
Each one has the evidence, root cause, end-to-end test result, and risk.
1
Factories/marketing
Waiting for you- asked just nowDay 14 of 141
Claims cited
48
each linked to its source
Posts published
5
through your CMS
Sources kept
3.8%
share of pages researched
Spend
$10.30
of $25.00 cost budget
Research by sourcelast 14 days
1,210pages researched
News412Industry reports286Research papers351Product docs97Earlier posts64
Agent tasksLatest first
| Task | Status |
|---|---|
| Check earlier posts for used angles | Finished |
| Research the topic on the web | Working |
| Publish the post on faster month-end close | Waiting for you- asked just now |
| Link a source for every claim | Working |
| Outline one new angle | Finished |
Run plan
Each factory run researches a topic, writes in your voice, then publishes the post through your CMS.
- Read earlier posts to pick a new angle
- Link every claim to its source
- Follow your CMS publish policy
- Cost budget $25.00, warn at 70%
Ready for your review
Publish the post on faster month-end close
Your CMS publish policy asks for approval. The post and its sources are ready.
Factories/sales
WorkingDay 14 of 14Potential customers
12511%
vs yesterday
Qualified
36
29% of found
Active agents
6
running now
Spend
$16.72
of $25.00 cost budget
New potential customers / dayDay 14 of 14
125found so far
Day 1Day 7Day 14
SourcesWhere they were found
125found
- Public web64
- X34
- Reddit18
- Forums9
Potential customersNewest first
| Profile | Source | Fit | Status |
|---|---|---|---|
| Ops lead, construction supplierAsked peers how they handle supplier credits | X | - | Prospecting |
| Finance manager, e-commerce brandReviewed an AP tool as too slow to set up | 80 | Qualified | |
| AP specialist, food importerHiring a second AP clerk | X | 71 | Qualified |
| CFO, 25-person agencyAsked on X for an invoice matching tool | Forums | 77 | Qualified |
| Bookkeeper, family-owned wholesalerShared a matching spreadsheet template | Public web | 62 | Prospecting |
Run plan
Read public Reddit and X posts, check Gmail, then draft replies for people who asked for what you sell.
- Read public web pages
- Read public Reddit and X posts
- Check Gmail for earlier threads
- Skip people already contacted
- Cost budget $25.00, warn at 70%
- Draft only. Never send.
Ready for your review
12 reply drafts in Gmail
Drafts for the top 12 are saved, not sent.
1
Factories/security
Waiting for you- asked just nowDay 14 of 141
Safe probes sent
1,284
read-only, inside your scope
Confirmed findings
3
1 high severity
Routes checked
41
new or changed, 38 clean
Spend
$4.00
of $25.00 cost budget
Probes by checklast 14 days
1,284probes sent3 confirmed
Missing auth312Exposed data486Verbose errors174Security headers208Open redirects104
Agent tasksLatest first
| Task | Status |
|---|---|
| Find the latest deployment | Finished |
| Probe new API routes for exposed data | Working |
| Review 3 draft patch pull requests | Waiting for you- asked just now |
| Redact tokens from the evidence | Finished |
| Write a regression test for the auth fix | Working |
Run plan
Each factory run probes new and changed routes in your latest deployment with safe, read-only requests, then patches each confirmed finding.
- Send only safe, read-only probes in scope
- Confirm each finding with request and response
- Open each patch as a draft pull request
- Cost budget $25.00, warn at 70%
Ready for your review
Review 3 draft patch pull requests
Each one has the redacted request and response, severity, fix, and test result.
1
Factories/finance
Waiting for you- asked just nowDay 14 of 141Passages compared
12,480
against the last digest
Material changes
17
met your materiality bar
Digests filed
9
dated pages in Notion
Spend
$3.46
of $25.00 cost budget
Source readslast 14 days
1,060pages read17 material changes
Agent tasksLatest first
| Task | Status |
|---|---|
| Read central bank news | Finished |
| Compare sources with the last digest | Working |
| Review today's digest of 3 changes | Waiting for you- asked just now |
| Estimate revenue at risk from Stripe | Working |
| Quote old and new text for each change | Finished |
Unleash the full power of AI
There's always work to do.
Run your own AI agent factory and make it your edge.
Agents at once One assistant1FactorAI100sArmy of agent swarms
Agent swarms split up the work and run every piece in parallel.
Hours of work per day Your team8hFactorAI24/701224Always running
Your agent swarms run 24/7 and keep working long after your team logs off.
Spend per run Usage billingNo capFactorAIYour capSpend you control
Set a dollar limit on every run, so you always know the most it can cost.
How it works
Build the factory that runs itself.
Scale horizontally. Same budget, more agents, more results.
Assign a task, and the factory plans it, runs it in parallel, and gets cheaper with every run.
Factories/Invoice matching
Idle0 agentsDescribe a problem worth solvingStart
Describe a problem to start a run
Agents work on it in parallel and wait for approval where you require it.
Agents0/12
Invoices checked0
Best so far-
Parallel runs0/12 done
- 01
- 02
- 03
- 04
- 05
- 06
- 07
- 08
- 09
- 10
- 11
- 12
Compared side by side2,400 samples
- Three-way match0%
- Line-item match0%
- Fuzzy totals0%
- Vendor rules0%
Ready for your review
Roll out Three-way match to every invoice
Matched 97.8% of 2,400 sample invoices.
Nothing rolled out. The current process stays as it is.
Your own agent factory
Hundreds of agents.
Working around the clock.
Give it a goal and a budget. FactorAI spins up the agents, stays on budget, and delivers verified work.
See the whole factory at work
Every run, step, and result in one view. Spot what is stuck or waiting for you.
Runs on your schedule
Set a cron or interval schedule once. Runs start on time, even overnight, and the results wait for you.
Scale the work. Keep control.
You set the permissions, tools, and limits. Allowed actions run. Actions that need approval wait for you.
Your model. Your keys. Your harness.
Pick a model per chat and a harness per team, project, or chat. Bring your own OpenAI, xAI, or OpenRouter key, or a ChatGPT subscription.
Enterprise Ready.
Roles, approvals, audit logs, and run limits are built in.
Access control
Set team roles and project roles for who can view, edit, and run work.
Custom approvals
Choose which actions are allowed, denied, or need a person to approve.
Audit visibility
Review what happened in the team audit log.
Policy boundaries
Set spend, time, and step limits for runs, and a monthly spend limit for the team.
Human oversight
Runs wait for a person whenever an action needs approval.
Self-hosting
Run in your own infrastructure with your choice of LLM providers. Contact sales for an Enterprise deployment.
Maximize customizability.
Integrates with your stack.
Bring your own keys, harness, and models. Connect them once, then work across multiple identities.
Everything runs through accounts you control
- Permissions
- Budgets
- Observability
- Approvals
Platform activity
A snapshot of activity across the platform.
- Tokens
- 602M
- AI messages
- 10.6K
- Graphics
- 592
- Messages sent
- 888
Pricing
Choose the monthly FTAI and support level that match the way you work.
What's included:
$1 daily login bonus
Everything in Hobby, plus:
Pay as you go
Multi-member teams
Free viewers
Beta Feature Opt-in
Build the future.
Scale to millions.
Start free. Daily FTAI plus a 7-day Pro trial.
Start building$
Launch from Claude Code, Codex, Cursor, and other agents.