Pricing
One subscription for every Philotic surface. Transparent usage, no opaque credits, no surprise overage.
All prices in USD. Annual billing saves 20%.
Free
Try the substrate. No credit card.
What you get
- Personal account, single device
- Limited daily agent actions
- Frontier models at fair-use rate
- Email support
Usage
~50 agent actions / day
Pro
The daily driver for one operator.
What you get
- Mobile + Desktop, multi-device sync
- Unlimited connectors (GSuite, Slack, GitHub, …)
- All Philotic-1 variants · auto / fast / reason
- Priority routing on frontier models
- X-Ray on every inference
Usage
2,500 agent actions / mo · transparent overage at $0.012
Team
For small teams running on the substrate.
What you get
- Everything in Pro, per seat
- Shared workspaces and agent libraries
- Bring-your-own keys (Anthropic, OpenAI, Google)
- Audit log · approval queue · role-based access
- Long-context reasoning on Philotic-1 · reason
- SAML SSO
Usage
6,000 agent actions / seat · pooled across team
Power
Heavy use. All surfaces unlocked.
What you get
- Everything in Team
- Auto-tune fine-tuning under operator approval
- Sovereign-style data residency (US / EU / UAE)
- Dedicated model capacity, no fair-use throttle
- Custom retrieval & connector SDKs
- Slack-channel support, business-hours SLA
Usage
20,000 agent actions / seat · pooled · soft cap
Transparent usage
You see what an agent run will cost before you start it, while it's running, and after it ends.
Every plan ships with the same usage meter the platform team uses internally. Actions are denominated in plain English, not opaque credits. Overage is billed at a fixed rate after a soft cap, and you can pin a hard cap for any seat.
No opaque credits
Everything is denominated in agent actions and tokens — read on the bill, on the dashboard, in the SDK response.
Predictable overage
Soft cap → fixed-rate overage → optional hard cap. No surprise bills, no rate-card hunting.
Pooled across team
On Team and Power, allocations pool across seats. One heavy week doesn’t starve another seat.
Auditable per inference
Every charge ties back to a trace_id. You can replay the run, see the chunks, and see why it cost what it cost.
Spend · MTD
$418.20
Projected
$612.00
Hard cap
$1,200
Recent run · trace 4f8e9b2c
Caliber RFP — 4-step research + draft
Compare plans
Every feature, every plan, in one table.
We don't hide capabilities behind a sales call. The only thing reserved for Enterprise is what genuinely takes deployment work — sovereign substrate, air-gap, custom contracts.
Surfaces
Models & inference
Agents & workflow
Trust & sovereignty
Support
One subscription, three surfaces
Same substrate. Mobile, Desktop, and Platform run on one identity, one trace, one bill.
Philotic Mobile
iOS · AndroidPersonal agentic management on your phone. Brief before meetings, draft replies, run quick missions on the go.
Philotic Desktop
macOS · WindowsThe same agents on your machine, plus the connector surface — GSuite, Slack, GitHub, Notion, your filesystem.
Philotic Platform
Web console · SDKThe substrate underneath. X-Ray, approval queue, fine-tuning, residency, customer-held keys. Self-serve on Team, full on Power, sovereign on Enterprise.
Enterprise
When you own the consequences, you should own the substrate. Sovereign deploy, customer-held keys, custom contracts.
Sovereign deploy
On your VPC, on-prem, or air-gapped. We ship the build, you hold the perimeter.
Customer-held keys
KMS-backed model encryption. Philotic never sees your weights or your data unkeyed.
Residency
Pin retrieval and inference to US, EU, or UAE. Manifest enforced at every request.
Compliance
SOC 2 Type II · ADSIC ND 133684 · UAE PDPL · GDPR. DPA on request.
Procurement
MSA, custom SLA, security questionnaire, redlines. We have done this before.
Pilots in 8 weeks
Typical sovereign deploy ships in eight weeks from signed MSA.
Why don’t you use opaque credits like other AI tools?
Because we watched users get burned by them. Credits hide what an action costs, and that breaks trust. Philotic bills in agent actions and tokens — both readable on every trace. You can predict an agent run before it starts.
What is an "agent action"?
One discrete unit of agent work — a model call, a tool invocation, or a retrieval. A typical 4-step research mission consumes ~200–400 actions. Every action is logged with a trace_id you can replay.
What happens if I exceed my plan’s allocation?
You hit the soft cap, the platform tells you, and overage is billed at a fixed published rate. You can pin a hard cap per seat or per workspace if you want a guarantee.
Do Mobile and Desktop need separate subscriptions?
No. One Philotic subscription covers Mobile, Desktop, and Platform across all your devices. Same identity, same trace, same bill.
Can I bring my own model keys?
Yes, on Team and above. Connect your own Anthropic, OpenAI, Google, or local provider keys. Philotic routes through them while preserving X-Ray trace.
How does fine-tuning fit into the price?
Auto-tune is included on Power and Enterprise. Each fine-tune is proposed with projected lift, the slice of weights touched, and a cost estimate, gated by approval.
Is my data used to train shared models?
No. Customer data never enters a shared training corpus. Fine-tunes you authorize live in your tenancy and never leave it.
Can I downgrade or cancel?
Anytime. Downgrades take effect at the end of the current billing period. Cancellation preserves read-only access to your traces and data exports for 90 days.
Do you offer startup or non-profit discounts?
Yes — startups under three years old and registered non-profits get 50% off Team for the first year. Contact us with details.