Get the most from your plan
Step-by-step setup for every tier — from spinning up your first agent on Free to governing a workforce on Business.
Free
Wire one project's agents into agentwach in under 10 minutes, with 10k events and 14-day retention.
Who it's for
Solo builders trying agentwach on a single side project or prototype before committing.
Setup in 4 steps
- 1Create your workspace
Sign up, name your workspace, and we'll seed a sample agent so the dashboard isn't empty.
- 2Add your first agent
Dashboard → Agents → New. Give it a name and pick an SDK snippet (Node, Python, or REST).
- 3Send your first event
Paste the snippet into your agent loop. The Live tab turns green the moment the first event lands.
- 4Set one budget
Open the agent → Guardrails → set a daily token cap. Even a loose cap catches runaway loops fast.
- →Free retention is 14 days — export anything you want to keep before it rolls off.
- →Stay under 3 agents and 10k events/month, otherwise ingest pauses until next month.
- →Use the CSV upload tab to backfill historical runs without writing code.
Hobby
Everything in Free plus token monitoring, 50k events, 30-day retention, and 10 agents.
Who it's for
Solo developers shipping side projects or indie SaaS that has real users hitting agents daily.
Setup in 4 steps
- 1Subscribe from Pricing
Pricing → Subscribe under Hobby. Checkout opens inline — no redirect.
- 2Turn on token monitoring
Dashboard → Tokens → Enable. We start tracking cost per agent, per model, per day.
- 3Add a token alarm
Tokens → New alarm. Pick threshold (e.g. $5/day) and an email destination.
- 4Scale to your real agents
Add up to 10 agents. Tag them by environment (prod/staging) so dashboards filter cleanly.
- →Email forwarding integration lets you ingest agent reports without API code.
- →Hobby keeps 30 days — long enough to debug a weekly incident postmortem.
- →Hitting 40k events/mo? Pro pays for itself once you cross ~80k.
Pro
Unlimited agents, 500k events, 90-day retention, priority support, and full API access.
Who it's for
Small teams running agents in production who need replay debugging and integrations.
Setup in 4 steps
- 1Invite your team
Workspace → Members → Invite. Roles: owner, admin, member. Up to 3 workspaces on Pro.
- 2Issue API keys
Settings → API Keys → Create. Scope per workspace. Rotate quarterly.
- 3Wire integrations
Connect Slack for alerts, OpenTelemetry traces, and CSV bulk upload for offline runs.
- 4Build a replay session
Open any incident agent → Replay tab → Step through tool calls and prompts frame-by-frame.
- →Pro unlocks the OTel traces endpoint at /api/public/otel/v1/traces — point any OTel SDK at it.
- →Use guardrails tab to define hard caps that hard-block agents before they burn budget.
- →Schedule weekly export reports from Reports → Export to keep a permanent audit trail.
Team
10 workspaces, 2M events, 1-year retention, SSO prep, and dedicated support.
Who it's for
Engineering teams running agents across multiple products or environments at scale.
Setup in 4 steps
- 1Plan your workspace topology
Typical split: prod / staging / per-product. Map this on a whiteboard before creating workspaces.
- 2Enable SSO prep
Settings → Auth → request SSO. Support emails you a config template within 1 business day.
- 3Set up multi-workspace alerting
Create alarm groups so a single incident pages on-call across workspaces instead of N times.
- 4Define retention policies
Reports → Retention. Pin compliance-critical agents to the full 1-year window; let others roll.
- →Team is the sweet spot if you have 3+ engineers pushing agent changes weekly.
- →Use the diagnose tab to attach root-cause notes to incidents — searchable across all workspaces.
- →Overage on Team kicks in at 2M events; budget for $5 per extra 100k.
Business
Everything in Team plus Organization with seats, per-worker agent assignments, and configurable token & cost limits with warn-or-block enforcement.
Who it's for
Companies that need to govern how individual employees use AI agents — with hard limits, per-user reporting, and assignable access.
Setup in 5 steps
- 1Create your organization
Org → New organization. Name it after your company. You become the owner automatically.
- 2Invite workers as seats
Org → Members → Invite. Pick a role: owner, admin, manager, or member. Invites expire in 7 days.
- 3Assign agents per worker
Org → Agents. Drag-assign which agents each employee can use — most workers only need 1–3.
- 4Set per-user policies
Org → Policies → New. Define token + cost caps per period, the warn threshold (e.g. 80%), and pick warn-only or hard-block on exceed.
- 5Review weekly usage
Org → Usage. Per-employee dashboards show tokens, cost, and which policies fired warnings.
- →Start with warn-only enforcement for two weeks to calibrate, then flip the most expensive agents to hard-block.
- →Org-wide policies (no agent set) act as a safety net; per-agent policies override them.
- →Use manager role for team leads — they can edit assignments but not billing.
Enterprise
SAML SSO, private deploy, custom retention, and a dedicated SLA on top of every Business feature.
Who it's for
Regulated industries and large orgs needing private infra, identity federation, and contractual guarantees.
Setup in 4 steps
- 1Book a scoping call
Contact → Enterprise. We'll align on identity provider, data residency, and SLA needs.
- 2Configure SAML SSO
We provision your IdP metadata, map groups → org roles, and validate with a pilot group.
- 3Provision private deploy (optional)
We deploy a dedicated agentwach instance in your cloud account or ours, with VPC peering.
- 4Sign the SLA
Choose 99.9% or 99.95% uptime, RTO/RPO targets, and dedicated support channel.
- →Most enterprises start on Business while SSO and procurement finalize — there's no lock-in to upgrade.
- →Ask about SCIM provisioning if your IdP supports it; it removes manual seat management entirely.
- →Custom retention can go up to 7 years — designed for teams preparing for SOC 2, HIPAA, or ISO 27001 workflows.