Skip to content

Pricing

Tether’s local engine stays source-available under BSL 1.1 and free for personal, academic, and commercial use that does not compete with FastCrest’s hosted offering.

Paid FastCrest products are for teams that need repeatable deployment proof: hosted GPU verification, signed proof packets, compliance evidence, rollout receipts, and enterprise support.

Open source

$0 / forever

Full local CLI for export, serve, replay, tracing, safety checks, and evidence generation.

Best for individual developers, labs, and teams proving feasibility on their own hardware.

Hosted proof job

$99 small / $249 large

FastCrest runs export + verify on hosted GPUs and returns a signed proof packet.

Includes:

  • Real export and parity verification
  • Latency receipt and deployment notes
  • Signed parity.cert.json + PARITY.md
  • Proof artifacts retained for 90 days

Team proof plan

$499 / month

For teams iterating on robot policies and needing repeated proof runs.

Includes:

  • 10 hosted proof jobs per month
  • Priority queue
  • Saved proof history
  • Compliance workspace add-on access
  • Email support

Enterprise

Custom

For regulated, audit-heavy, or fleet-scale deployments.

  • On-prem deployment (no HF Hub dependency)
  • SSO (SAML / OIDC)
  • Dedicated support engineer
  • SLAs on response time and proof turnaround
  • Custom embodiments and deployment gates
  • Continuous self-distillation or training-loop integration

Contact us — typically responds in 1-2 business days.

CapabilityOpen sourceHosted proof jobTeam proof planEnterprise
tether go / tether serveyesartifact inputartifact inputyes
Local export + replayyesoptional inputyesyes
Hosted GPU verificationone job10 jobs/monthcustom capacity
Signed parity certificatelocal onlyyesyesyes
Latency receipt / deployment noteslocal onlyyesyesyes
Proof historylocal files90-day retentionsaved workspacecustom retention
Compliance workspace add-onoptionalavailableincluded/custom
Continuous self-distillationmanualdesign partnercustom
Supportcommunityemailpriority emaildedicated channel
LicenseBSL 1.1service termsservice termsnegotiated
  • Hosted GPU verification run by FastCrest
  • Long-term artifact retention beyond the included job window
  • Compliance workspace collaboration
  • Enterprise SLAs, on-prem packaging, and custom deployment gates
  • Hardware, unless separately contracted

For teams that want help getting the first proof run working, FastCrest can run a fixed-scope deployment proof sprint.

Typical sprint inputs:

  • Model family or checkpoint
  • Target hardware
  • Replay/session data
  • Latency budget
  • Safety guard config or robot limits
  • Pass/hold criteria

Typical sprint output:

  • Proof packet
  • Latency and deadline-miss receipt
  • Export/parity certificate
  • Deployment blockers
  • Recommended next actions
  • Hobbyists / students / hackathon teams — Open source. Always.
  • Academic labs publishing on VLAs — Open source for experiments; hosted proof jobs for reproducible deployment artifacts.
  • Robotics startups (5-50 people) — Team proof plan once policy updates need repeatable release evidence.
  • Robotics OEMs / integrators (50+ people) — Enterprise for custom gates, on-prem, retention, and SLAs.

Use the hosted Cloud flow for proof jobs, or contact FastCrest for a proof sprint, team workspace, or enterprise deployment.