Saddler is now part of Beta - bringing AI work loops to teams.
Describe work once.
Run it reliably.
A managed workbench for tasks, playbooks, approvals, browser runs, outputs, and reusable learning.
No credit card. 200 free runs. Bring Your Own Provider or use ours.
See what Saddler can build
Every example shows expected cost and duration before you spend credits.
01 Examples
Pick a task. Inspect a result.
Open a teardown to see inputs, action highlights, files, cost, limits, and what we would change next.
Lead Enrichment Playbook
23m 48s run, $8.82 cost, 3 artifacts.
Enriched CSV, confidence brief, source log.
Open exampleStock Analysis Playbook
1 report bundle, source log, model posture visible.
Assumptions memo, sensitivity table, report bundle.
Open exampleResearch Brief Playbook
Citations table plus known-gaps result.
Brief, cited source table, gaps list.
Ticket Triage Playbook
Review step before status changes.
Prioritized queue, owner suggestions, risk notes.
Invoice Reconciliation Playbook
Temporary upload, 11 exceptions, approval result.
Exception workbook, audit notes, payment hold review.
02 Proof
See one run before you read another promise.
This sample result shows the source, output, cost, duration, and the approval that stops the playbook before it changes a live system.
Lead enrichment run
42 accounts in. 3 usable outputs out.
42 records checked, 38 high confidence, 4 left as draft-only.
- Source
- 42-row target CSV plus approved public sources
- Output
- Enriched CSV, summary, source log
- Cost
- $8.82
- Duration
- 23m 48s
- Approval
- CRM write plan paused before writeback
- Saved
- About 3 hours of research and CRM prep
What happened
- Validated 42 input rows and separated 4 records with missing company domains.
- Read CRM account fields in view-only mode before proposing any writeback.
- Searched public sources and attached citations to high-confidence enrichment fields.
Outputs
Decision waiting
Approve the CRM write plan for 38 high-confidence rows, deny it, or request changes before anything writes back.
Open the teardown03 How it works
A reusable playbook leaves a source trail.
The product stays simple when every output keeps the work that produced it.
Library
Reusable context, skills, memory, startup bundles, and connections.
Playbook
A reusable workflow for repeatable tasks.
Run
One recorded execution with inputs, actions, approvals, usage, and diagnostics.
Output
A produced file with the run and playbook that created it.
04 Two modes
Start personally. Add governance when the work becomes shared.
For individuals
You get
- Built-in credits
- Personal Library
- Playbook versions
- Run outputs and results
What you give up
No member management, workspace approval policy, or audit log.
For teams
You get
- Roles and invitations
- Shared Library
- Bring Your Own Provider custody
- Budgets, approvals, and audit
What you take on
More setup before playbooks can act on shared systems.
05 Trust
Trust comes from seeing the work.
Before an agent changes anything important, you should know what it used, what it did, what it cost, and what still needs a person.
Where does my data live?
Workspace-scoped records stay in the authenticated console. Public examples show only display-safe run anatomy.
Who pays for model calls?
Run usage shows cost and provider posture. Use Saddler-managed credits or Bring Your Own Provider.
Can I see what an agent did?
Run Detail records inputs, action log, approvals, produced artifacts, usage, and diagnostics.
Can I revoke access?
Setup handles readiness and repair. Organization handles governance for team workspaces.
06 Pricing
Plans describe the operating mode, not a feature maze.
Free
$0
- 200 free runs
- Starter templates
- Personal Library
- Run outputs with source context
Solo
$19 / mo
- More included runs
- Versioned agents
- Bring Your Own Provider
- Run history and artifact search
Team
$49 / seat / mo
- Shared Library
- Approval policy
- Budget posture
- Audit and owner controls
What a run costs
Operators need the cost of work, not just a seat price.