Future AGI pricing: what are the exact limits for Starter vs Growth (seats, projects, traces/month, retention)?

Most teams looking at Future AGI pricing want a crisp view of what you actually get in the Starter vs Growth plans: seats, projects, traces/month, and data retention. Below is a clear, engineering-level breakdown so you can decide which plan fits your current evaluation and monitoring workload.

Note: This article focuses on Starter vs Growth. Enterprise is fully custom (usage, support, deployment, and pricing) and is best discussed directly with the team.

---

Quick Answer: Starter vs Growth at a Glance

• Starter (Free, $0 forever)

  • Up to 3 team members
  • 3 projects
  • 10K traces/month
  • 1 GB storage
  • Basic reporting, community support
  • Historical lookback & full-resolution data retention: 120 days

• Growth (Paid, pay-as-you-scale)

  • Unlimited team members
  • More projects and experiments (higher limits than Starter; contact Future AGI for exact caps as they evolve)
  • Starts at 100K+ traces/month and scales with usage
  • 10 GB storage
  • Advanced analytics, priority support
  • Historical lookback & full-resolution data retention: 180–360 days depending on current configuration

Because Growth is usage-based and optimized for scaling teams, some caps (like trace volume beyond the initial tier) are dynamic and configured with your account representative.

---

The Quick Overview

• What It Is: A pricing model designed to let you start evaluating and monitoring LLM agents at $0, then scale up to high-volume, production-grade monitoring and evaluation with pay-as-you-go usage.
• Who It Is For:
  • Starter: Small teams, prototypes, hackathons, early-stage startups validating their agent architecture.
  • Growth: Teams running real workloads (RAG chatbots, voice agents, internal copilots) that need serious trace volume, better analytics, and longer retention.
• Core Problem Solved: LLM apps are probabilistic and inconsistent. You need traces, evaluations, and retention to reproduce failures, compare prompts/models, and protect production traffic. The pricing tiers match that lifecycle.

---

How the Pricing Structure Works

Future AGI pricing ties directly to how much you evaluate and observe your agents:

• Seats & Projects determine how many people and workflows you can coordinate.
• Traces/month determines how many conversations, tool calls, and workflow runs you can observe and debug.
• Retention windows determine how far back you can replay failures, run retroactive evals, and feed production data into Improve and Monitor & Protect.

At a high level:

1. Starter – Validate the Workflow
   • Ship early experiments, wire up tracing, and get a feel for the Datasets → Experiment → Evaluate → Improve loop without paying a cent.
2. Growth – Scale to Real Traffic
   • Move from demos to production-like volumes with more traces, longer retention, advanced analytics, and priority support to keep reliability high as you scale.
3. Enterprise – Customize for Heavy Production
   • For teams that need on-prem, custom SLAs, and very high-volume traces/evals, pricing and limits are negotiated to match infra and compliance needs.

---

Starter Plan: Exact Limits and What They Mean

Future AGI’s Starter plan is designed to be “$0 forever (seriously)” while still letting you run real evaluation workflows.

Seats (Team Members)

• Limit: Up to 3 team members
• Implication:
  • Ideal for a small squad: 1–2 ML/AI engineers + 1 PM or QA.
  • Enough to collaborate on datasets, experiments, and eval dashboards, but intentionally constrained to keep it focused on early stagework.

Projects

• Limit: 3 projects
• Implication:
  • Think of a “project” as one cohesive LLM application or agent workflow (e.g., support chatbot, sales summarizer, internal research assistant).
  • With 3 projects you can:
    • Run a main production idea.
    • Keep a sandbox area for experimentation.
    • Have one more for a second use case or a dedicated “eval lab.”

Traces per Month

• Limit: 10K traces/month
• What is a trace?
  A trace is a logged execution of your AI workflow—covering messages, model calls, tool invocations, and intermediate steps.
• Implication:
  • 10K traces/month is enough for:
    • Dev and staging traffic.
    • Low-volume production pilots.
    • CI-style evaluation runs on every deploy.
  • If you have a high-traffic chatbot or voice agent, you’ll quickly hit this and should plan to move to Growth.

Storage & Data Retention

• Storage: 1 GB
• Historical lookback: 120 days
• Full-resolution data retention: 120 days
• Implication:
  • You can replay and analyze traces for up to ~4 months.
  • Suitable for:
    • Short iteration cycles.
    • Quarterly evaluations and audits.
  • For long-lived deployments where you want to compare behavior across half-year or annual cycles, Growth or Enterprise will be more appropriate.

Analytics & Support

• Reporting: Basic reporting (the essentials)
  – Core metrics and dashboards to understand failure modes and performance.
• Support channel: Community support
  – Access to docs, community, and general help, but not a dedicated response-time SLA.

---

Growth Plan: Limits and Scaling Levers

The Growth plan is designed to “accelerate your growth with tools that actually work” and is usage-based. It keeps friction low while you move from prototype to production scale.

Seats (Team Members)

• Limit: Unlimited team members
• Implication:
  • Enable broader collaboration:
    • Multiple product squads.
    • Data/ML teams.
    • Quality and safety reviewers.
  • Centralize evaluation and tracing across the org without worrying about seat caps.

Projects & Experiments

From the internal plan comparison and UI guidance:

• Projects: Higher than Starter; effectively more projects for Growth (exact numeric cap is subject to change and can be confirmed in-app or with sales).
• Experiments: More than Starter (Starter shows 3 experiments; Growth shows 5+ and often trending to higher or unlimited for active accounts).
• Implication:
  • Enough verticals for:
    • Multiple live apps (support, sales, search, analytics).
    • Dedicated research projects.
    • Separate environments per business unit.
  • More concurrent experiments so you can A/B/C compare prompts, models, tools, and routing policies without constantly deleting old runs.

Traces per Month

Pricing tables reference:

• 100K traces/month tier
• Traces beyond base volume at ~$8–10 per 100K traces (pay-as-you-go)

Concretely:

• Starting point: Expect at least 100K traces/month as a baseline for Growth.
• Scaling: You then scale usage up via pay-as-you-go:
  • Traces: ~$8–10 per 100K traces depending on the latest pricing.
• Implication:
  • Growth is built for:
    • Production chatbots.
    • Live agents with real user traffic.
    • CI/CD evaluation runs over large synthetic datasets.
  • You can track and evaluate millions of interactions/month by scaling trace volume.

Storage & Data Retention

From the pricing and comparison data:

• Storage: 10 GB (base)
• Historical lookback: 180–360 days (documentation notes 360 days in some comparison sections)
• Full-resolution data retention: 180–360 days
• Implication:
  • You can track behavior across half-year+ cycles, which is crucial for:
    • Seasonality.
    • Model/regression drift.
    • Longitudinal safety monitoring (toxicity, PII, prompt injection trends).
  • 10 GB storage is enough for high-volume trace logs, depending on trace payload size; you can purchase additional storage as you scale.

Advanced Features & Support

Growth adds production-oriented capabilities:

• Advanced analytics:
  • Deeper dashboards, drill-downs, anomaly detection, outlier detection, and Watchdog automated insights.
  • Easier to “pin-point root cause” across workflows and model variants.
• Advanced workflows:
  • Evals in CI/CD – hook evaluations into your deploy pipeline.
  • AGI x and complex workflows – more support for agentic chains, multi-tool orchestrations, and structured experiments.
• Support channel: Email with priority support
  • Faster response times than Starter; business-hours SLAs apply for Growth.
• Usage-based pricing:
  • Evaluations: ~$0.002–0.01/API.
  • Storage: ~$5/GB beyond included.
  • Synthetic data: ~$100/1M tokens.
  • Traces: ~$8/100K traces.

---

Which Plan Should You Choose?

When Starter Is Enough

Use Starter if:

• You have up to 3 team members actively working on the agent.
• You’re focused on prototyping:
  • Building your first RAG chatbot.
  • Defining evaluation metrics.
  • Testing a voice agent or internal summarizer.
• You only need:
  • 3 projects
  • 10K traces/month
  • 120 days of data retention
• You’re comfortable using community support while you learn the platform.

This is where you validate:
Can we connect our app, generate synthetic datasets, run a few experiments, and see measurable quality gains?

When Growth Is the Right Move

Move to Growth when:

• You have more than 3 people across engineering, data, and product who need access to evaluation and tracing.
• You’re running production or near-production workloads, where:
  • 10K traces/month is no longer close to enough.
  • You want to plug evaluations into CI/CD.
  • You need longer retention (180–360 days) to analyze regressions and model changes over time.
• You care about:
  • Advanced analytics to debug multimodal agents.
  • Priority support to unblock production issues quickly.
  • Pay-as-you-go flexibility for traces, evals, and synthetic data.

If your AI systems already have real users, you’ll almost certainly want Growth so you can monitor and protect reliably.

---

FAQs: Limits, Overages, and Upgrades

What happens if I hit the 10K trace limit on Starter?

Short Answer: You’ll need to upgrade to Growth to keep logging new traces in that billing period.

Details:
The Starter plan is intended for low-volume usage. Once you approach 10K traces/month, new traces may be rate-limited or blocked until the next monthly cycle. To avoid blind spots in monitoring:

• Monitor your trace usage in the dashboard.
• Upgrade to Growth before you hit the cap if you’re pushing toward production volumes.

---

Can I keep Starter forever, or is there a time limit?

Short Answer: Starter is $0 forever—no forced upgrade.

Details:
Future AGI’s Starter tier is designed to give you a persistent, free environment to:

• Experiment with new agents.
• Run small-scale evaluations.
• Keep a sandbox even if your main workloads move to Growth or Enterprise.

You can upgrade to Growth for production workloads and still maintain a Starter-like setup in parallel for smaller projects or experiments (subject to org configuration).

---

How is data retention different between Starter and Growth?

Short Answer: Starter gives you ~120 days; Growth and above give you 180–360 days.

Details:

• Starter:

  • Historical lookback: 120 days
  • Full-resolution data retention: 120 days

• Growth:

  • Historical lookback: typically 180–360 days
  • Full-resolution retention: 180–360 days

Longer retention means you can:

• Replay incidents that happened many months ago.
• Compare behavior across model versions (e.g., GPT-4 vs GPT-4.1 vs custom finetunes).
• Perform long-range trend analysis on safety metrics (PII, toxicity, prompt injection).

For even longer retention or strict regulatory requirements, Enterprise plans can support custom retention policies and on-prem options.

---

How do synthetic data and evaluations factor into pricing?

Short Answer: They’re usage-based add-ons, especially relevant for Growth and Enterprise.

Details:

• Synthetic data: ~$100 per 1M tokens generated.
• Evaluations: ~$0.002–0.01 per evaluation API call.
• Traces: ~$8–10 per 100K traces (beyond included volume).

In practice:

• Starter teams often stay within built-in limits.
• Growth teams treat synthetic data and eval calls as part of a CI-style test budget—similar to paying for unit/integration test infra.

---

Summary

Future AGI’s pricing is designed to match the lifecycle of real AI products:

• Starter gives you:
  • Up to 3 teammates, 3 projects, 10K traces/month, and 120 days of retention.
  • Enough to build your Datasets → Experiment → Evaluate → Improve loop and wire up traces without any cost.
• Growth unlocks:
  • Unlimited seats, more projects and experiments, 100K+ traces/month starting volume, 10 GB storage, and 180–360 days retention.
  • Advanced analytics, priority support, and pay-as-you-go scaling for evals, traces, and synthetic data.

If you’re still in prototype mode, Starter is the right baseline. As soon as your agents see real traffic—and you care about long-term reliability, safety, and CI-style evaluation—you’ll want Growth.

---

Next Step

Ready to figure out whether Starter or Growth is the right fit for your current workloads—or to get exact limit details for your use case?

Get Started