Pricing
Start free. Pay for what you need to approve it.
The Open SDK is free, self-hosted, and yours to read line by line. Everything above it is the shared intelligence that keeps building, testing, and governing your agents underneath — and every number on this page is pulled from the same rate table the platform actually bills against.
Open
Individual engineer, no procurement
$0 — free
+5,000 welcome tokens on signup
- Vouchstone Open SDK + standalone harness runtime
- Self-hosted, bring your own LLM keys
- Command-line tools to build and test agents
- No credit card or account required to start
Platform
Team self-serving past free tier
$59/seat/mo + metered overage
Minimum 50 seats · +2,000 tokens/seat/mo included
- Hosted extraction pipeline, self-serve KG wizard
- Wiki, Company Brain + Council
- Agent Studio, Agent Harness, Evals, and Guardrails included
- Monthly token allowance per seat, then metered overage
Enterprise
Security/compliance-gated rollout
$59/seat/mo + $15,000/agent/yr
Minimum 50 seats · +2,000 tokens/seat/mo included
- Everything in Platform
- Agent Optimization tuning, custom rules for agent behavior, full autonomy control
- Fine-grained access controls, dedicated budget, exportable audit trail
- Dedicated support team
Sovereign & Air-Gapped
Highly regulated industries
Contact us
Custom annual license — priced per engagement
- Everything in Enterprise
- Signed, fully self-hosted deployment package
- Turnkey self-hosted LLM inference — no external API calls, ever
- Custom annual license, scoped and quoted per engagement
Implementation acceleration and messy-data-estate modeling are available as an optional services add-on layered on any paid tier — never a required entry point. Not sure which tier fits? Run the free pre-assessment for a computed recommendation.
Full metered rate table
Platform and Enterprise seats include 2,000 tokens/seat/mo — usage beyond that draws from this same rate table (1 USD = 100 Vouchstone tokens). This is the exact same rate table the backend bills against, not a marketing approximation.
| Operation | Cost (Vouchstone tokens) | ≈ USD (100 tokens = $1) |
|---|---|---|
| Agent execution — simple chat | 15 | $0.15 |
| Agent execution — complex (tools/RAG) | 75 | $0.75 |
| Agent execution — advanced multi-step reasoning | 150 | $1.50 |
| Create an agent | Free | — |
| Update an agent | Free | — |
| Workflow run — base | 25 | $0.25 |
| Workflow — additional step | 10 | $0.10 |
| Workflow run — complex (conditionals) | 100 | $1.00 |
| Create a workflow | Free | — |
| Update a workflow | Free | — |
| Collective coordination — base | 75 | $0.75 |
| Collective — additional agent | 25 | $0.25 |
| Collective — consensus mechanism | 50 | $0.50 |
| Memory query (semantic search/retrieval) | 5 | $0.05 |
| Memory store — per 100 vectors | 10 | $0.10 |
| Memory store — per 1K vectors | 50 | $0.50 |
| Memory — advanced vector search | 8 | $0.08 |
| Memory consolidation job | 25 | $0.25 |
| Document ingested | 20 | $0.20 |
| Document chunk created | 5 | $0.05 |
| Document indexed | 15 | $0.15 |
| External API call | 2 | $0.02 |
| Integration sync (Slack, JIRA, etc.) | 10 | $0.10 |
| Integration event received | 5 | $0.05 |
| Webhook trigger | 3 | $0.03 |
| LLM input — per 1K tokens | 2 | $0.02 |
| LLM output — per 1K tokens | 6 | $0.06 |
| LLM embedding — per 1K tokens | 1 | $0.01 |
| Decision graph traversal | 30 | $0.30 |
| Decision graph — node evaluated | 5 | $0.05 |
| Decision graph — conditional branch | 8 | $0.08 |
| Optimization Studio — DSPy run | 250 | $2.50 |
| Fine-tuning job | 1,000 | $10.00 |
| Fine-tuning job — large model | 2,000 | $20.00 |
| A/B test variant | 50 | $0.50 |
| Storage — per GB/month | 25 | $0.25 |
| Bandwidth — per GB | 10 | $0.10 |
Show your work
Every quote from the pre-assessment is computed from the same public formula — no negotiated black box. Here it is worked through for one illustrative scenario: a compliance-evidence rollout across Jira, Confluence, GitHub, and Slack, 12,000 documents, 2,500 governed agent actions a month.
Recommended tier
platform
Computed price (monthly)
$2,413 – $2,975
Every dollar figure on this page traces back to a real backend constant served by GET /api/v1/billing/rates or GET /api/v1/pre-assessment/example — nothing here is hardcoded in the frontend, and a backend test asserts these numbers can never silently drift from what the platform actually bills.