Control Spend Before Agents Multiply It
Traditional AI billing tools were designed for simple prompt-and-response usage. ModelSpend is built for the next wave: autonomous agents, multi-step workflows, coding assistants, MCP tools, background jobs, and AI systems that can generate dozens or hundreds of model calls from a single user request.
The Founding Beta is your opportunity to evaluate ModelSpend on your real agentic workloads before the platform is generally available — and to shape the roadmap as a founding participant.
From one request to many calls
Traditional AI billing tools were designed for simple prompt-and-response usage. ModelSpend is built for the next wave: autonomous agents, multi-step workflows, coding assistants, MCP tools, background jobs, and AI systems that can generate dozens or hundreds of model calls from a single user request.
One user request can trigger dozens of downstream model, tool, and MCP calls.
ModelSpend adds routing, policy, budgets, and visibility across the entire chain.
Use ModelSpend During the Beta Period Without Payment
Founding Beta participants receive free access until the launch promotion countdown ends. This gives you time to test ModelSpend with real prompts, routing policies, budgets, dashboards, and support workflows before deciding whether it fits your team.
Choose Your Founding Beta Path
Start individually, bring a team, or apply for Enterprise Design Partner access if you are evaluating agentic AI governance at scale.
For individual developers, founders and operators evaluating AI routing and spend control.
- Personal beta access
- API keys and routing tests
- Dashboard visibility
- Feedback through support tickets
For small teams testing AI usage, routing policies, budgets and reporting together.
- Team-oriented evaluation
- Policy and budget workflows
- Usage and cost visibility
- Support feedback loop
For businesses assessing AI governance, reporting and cost control across real workloads.
- Business evaluation access
- Governance and routing review
- Audit and reporting workflows
- Roadmap feedback priority
Important Participation Conditions
Pre-release software; features may change, be withdrawn, or become temporarily unavailable.
Outputs, cost estimates, routing decisions, forecasts, alerts, and recommendations may be incomplete, delayed, or inaccurate.
Not suitable as the sole basis for financial, operational, legal, security, compliance, procurement, employment, medical, psychological, safety-critical, or other business-critical decisions.
Free access ends when the campaign countdown reaches zero.
Inactive accounts may lose beta access after 30 days of non-use.
Enterprise Design Partner access requires separate review and may require additional confidentiality terms and onboarding checks.
ModelSpend may suspend or withdraw access for misuse, security risk, inactivity, breach of terms, or campaign closure.
Full Beta Terms are available at /founding-beta/beta-terms/. Participation in the beta is subject to those terms.
Founding Beta FAQ
Common questions about free beta access, agentic workload support, feedback, and participation terms.
How long does free access last?
Free access lasts until the launch promotion countdown ends. The campaign is time-limited, so participants who join later may receive a shorter free access period than those who join early.
Is payment required?
No payment is required during the beta period. There is no automatic paid conversion at the end of the beta — if paid plans become available after the beta period ends, you will be informed and given a clear choice.
What are agentic workloads?
Agentic workloads are AI workflows where an assistant, agent, or automated system can make multiple model calls, tool calls, API requests, or decisions from a single user instruction. These workflows can be powerful, but they can also make AI usage and cost harder to predict and govern.
Why does agentic AI need cost governance?
Agentic systems can trigger repeated model calls, retries, tool usage, background tasks, or multi-step reasoning chains. Without routing, budgets, approvals, and visibility, costs can grow faster than expected — and can be difficult to attribute or audit after the fact.
Does ModelSpend only track human chat usage?
No. ModelSpend is designed to support both human-initiated AI usage and agentic workloads, including workflows that route through agents, tools, model providers, and automated systems.
Can ModelSpend help with MCP-connected tools and agents?
ModelSpend is built with MCP-connected and tool-calling AI workflows in mind. If you are evaluating MCP-connected workloads as part of your Founding Beta participation, please share your feedback through the support ticket system so we can prioritise production hardening.
Can I get enterprise access immediately?
Enterprise Design Partner access is available only to selected organisations and is not self-serve. Applications are reviewed by our team, and access may require additional confidentiality terms and onboarding checks.
What happens if I do not use my account?
Beta places are limited. If a participant does not make reasonable use of their account within the first 30 days, ModelSpend may withdraw beta access and make the place available to another applicant. Reasonable use includes logging in, connecting an integration, creating a routing policy, running test usage, or submitting feedback.
What feedback is expected?
Participants are asked to report bugs, confusing behaviour, errors, or improvement suggestions through the support ticket system in the platform. Your feedback directly shapes what we prioritise.
Is the beta suitable for business-critical decisions?
The beta is for evaluation and testing. You should not rely on it as the sole basis for financial, operational, legal, security, compliance, or other business-critical decisions. See the Beta Terms for full details.
Founding Beta Places Are Limited
If you are building with AI agents, coding assistants, tool-calling workflows, MCP-connected tools, or multi-step model pipelines, ModelSpend is built for the way your AI usage actually behaves.
Apply for Founding Beta Access