.png&w=3840&q=75)
FastRouter vs TrueFoundry: Pricing Compared
FastRouter vs TrueFoundry pricing, side by side. See what you pay for LLM gateway access and where costs can add up.

.png&w=3840&q=75)
The decision you're actually making
Both FastRouter and TrueFoundry offer a managed gateway, and both can meet a self-hosted or on-prem requirement through an Enterprise conversation. That's not what separates them.
What separates them is how each prices access:
- FastRouter charges a flat, predictable platform fee and passes tokens through at zero markup.
- TrueFoundry prices access to capacity in request-volume tiers, and doesn't publish per-token rates.
If deployment control is a hard requirement, put both vendors through the same Enterprise conversation rather than assuming only one can meet it. But for most teams evaluating these two, the decision comes down to pricing and scaling, not deployment model, and the pricing curves here are not close once you hit real traffic.
Published pricing (and what the numbers imply in production)
Here's the side-by-side pricing exactly as stated, followed by the “so what” for each line item. Because in production, the caps and limits are the product.
FastRouter | TrueFoundry | |
|---|---|---|
Free tier | Starter, $0/mo, 1M requests/mo, 1 BYOK key, 200K logs, 7-day retention | Developer, $0/mo, 50K requests/mo, 3 users, 7-day trial only |
Mid tier | Pro, $199/mo, 5M requests/mo, 10 BYOK keys, 1M logs, 60-day retention | Pro, $499/mo, 1M requests/mo, additional 2M requests costs another $499/mo |
Upper tier | Business, $799/mo, 50M requests/mo, 50 BYOK keys, 5M logs, 120-day retention | Pro Plus, $2,999/mo, 1M requests/mo (same request ceiling as Pro, priced for priority support/SLA) |
Enterprise | Custom, unlimited orgs/members/keys, volume discounts | Custom, 10M+ requests/mo, VPC/on-prem/air-gapped options |
Token cost | Zero markup, you pay the provider's listed token price, always | Not published per-token; billing is request-based, not token-based |
Self-hosting | Not the model; FastRouter is a managed gateway; a dedicated/on-prem path is available through Enterprise | Available; self-hosted infra runs roughly $600–$1,000/mo on top of the platform fee |
Free tier: this tells you how each company expects engineers to adopt
- FastRouter Starter ($0/mo) gives 1M requests/mo. That's not a toy allowance. It's enough to run a meaningful staging environment, or a small-but-real production feature, without immediately playing “which requests do we drop this week.”
- TrueFoundry Developer ($0/mo) gives 50K requests/mo and is a 7-day trial only. That's an evaluation window, not a usable free tier. If your procurement and security review process takes longer than a week (it will), you're either paying or stopping.
Also note the operational limits FastRouter makes explicit even on free:
- 1 BYOK key: you can wire it up, but you'll feel friction if you want separate keys per environment/team.
- 200K logs, 7-day retention: enough to debug a fresh integration, not enough to analyze regressions from “the model change we made last month.”
TrueFoundry's Developer tier does publish a log retention window — 7 days, same as FastRouter Starter's. Where it diverges is governance: Developer only ships “built-in” guardrails, with no RBAC, SSO, or audit logs — those unlock at Pro and Enterprise. FastRouter's governance basics (RBAC, project and key limits, input/output guardrails) are on Starter from day one.
Mid tier: this is where most teams land first — and where the economics diverge
- FastRouter Pro is $199/mo for 5M requests/mo. That's a tier you can often set and forget unless you're doing very high-throughput workloads.
- TrueFoundry Pro is $499/mo for 1M requests/mo, and “additional 2M requests costs another $499/mo.”
This structure matters. FastRouter says: pick a tier, get a lot of headroom. TrueFoundry says: capacity is a core billing unit; growth is something you pay for in chunks.
If you've ever been on call for a system that spiked traffic because of a launch, a retry loop, or a customer enabling a feature globally, you know why “chunks” are scary. Tier cliffs show up as either:
- throttling and degraded user experience, or
- a surprise bill, or
- an urgent procurement escalation to increase limits.
Upper tier: support/SLA vs capacity is the point
Two very different signals here:
- FastRouter Business is $799/mo for 50M requests/mo, plus higher BYOK/log/retention caps.
- TrueFoundry Pro Plus is $2,999/mo and still capped at 1M requests/mo, explicitly “priced for priority support/SLA.”
So Pro Plus is not a scaling tier. It's a “we need contractual support” tier.
That's not inherently bad. Plenty of orgs need SLAs and escalation paths. But you should not confuse “more expensive plan” with “more throughput.” In TrueFoundry's published tiers, paying more does not automatically buy more request headroom.
Enterprise: the categories split cleanly
- FastRouter Enterprise is Custom with “unlimited orgs/members/keys” and “volume discounts,” and this is also where FastRouter's own on-prem/dedicated deployment option lives — it isn't on the general pricing page, so it's easy to miss.
- TrueFoundry Enterprise is Custom for “10M+ requests/mo” and includes “VPC/on-prem/air-gapped options.”
So if you need self-hosted, VPC, or air-gapped deployment, both vendors can get you there — just through the same kind of custom Enterprise conversation, not through a published self-serve tier. TrueFoundry's published self-hosted infra cost ($600–$1,000/mo on top of platform fee) is a useful number to hold up against whatever FastRouter's Enterprise team quotes for the equivalent scope — ask for that number directly rather than assuming one vendor is the only option.
Token cost: this is a structural pricing difference, not a footnote
- FastRouter: “Zero markup, you pay the provider's listed token price, always.” That's clean. Your variable spend stays where it belongs: with the model providers. The gateway bill is the gateway bill.
- TrueFoundry: token pricing not published; billing is request-based, not token-based. That means the platform fee scales with a unit (“request”) that does not map cleanly to LLM cost in many real systems.
If your “request” is always the same shape (same prompt size, same output size), request-based billing can be predictable. If your system has heterogeneous requests (RAG vs chat vs extraction vs agent steps), “request” is a sloppy unit. One request can be tiny, the next can be huge. Same platform billing unit.
That's not a theoretical gripe. It changes how confidently you can forecast costs as your product evolves.
Self-hosting: TrueFoundry's biggest published differentiator, and a real tax
- FastRouter: managed gateway by default; a dedicated/on-prem path exists but is scoped through Enterprise, not published with a price.
- TrueFoundry: self-hosting available on a published path; self-hosted infra runs roughly $600–$1,000/mo on top of the platform fee.
That $600–$1,000/mo isn't just money. It's an operational commitment: you own uptime, upgrades, scaling, and incident response for the gateway layer. Some orgs are fine with that (or already staffed for it). Others underestimate it until the first “why is routing failing only in prod” incident.
If self-hosting/air-gapped is required, get both vendors' numbers before you sign up for that tax. If it's not required, don't casually take it on “just because.”
Where the gap actually shows up: normalize to cost per request (platform fee only)
Do not compare gateways by plan names. Compare them by what you're buying: platform overhead per unit of throughput.
- FastRouter Pro: $199 for 5,000,000 requests, roughly $0.00004 per request in platform fee, plus zero token markup.
- TrueFoundry Pro: $499 for 1,000,000 requests, roughly $0.0005 per request in platform fee — about 12.5x higher per request than FastRouter Pro, before token costs are even factored in.
That ~12.5x gap is the headline if you are cost-sensitive at scale. It's the difference between:
- “gateway overhead is background noise,” and
- “gateway overhead is something finance will notice.”
And because FastRouter is explicitly “zero markup” on tokens, you don't have a second hidden multiplier on the largest part of most LLM bills. You pay provider pricing for tokens; you pay FastRouter for platform.
TrueFoundry is “request-based, not token-based,” and does not publish per-token pricing. So the only thing you can say from published information is: the unit economics of the platform fee per request are materially higher at the comparable tier, and token pricing isn't transparent from what's public.
The weirdest TrueFoundry tier: Pro Plus is expensive and doesn't raise the cap
This is the part that will trip people up in procurement if they don't notice it early:
- TrueFoundry Pro Plus costs $2,999/mo
- and it's still capped at 1M requests/mo (same as Pro).
So when your team grows and you start needing faster support response or an SLA, you can pay a lot more without increasing throughput capacity at all.
That's not “bad pricing.” It's a very specific product packaging choice: support and contractual guarantees are priced separately from throughput.
Just don't buy it expecting headroom. It's not for headroom.
FastRouter's Business tier vs TrueFoundry's published ceiling: a scaling philosophy difference
- FastRouter Business ($799/mo) covers 50M requests/month.
- TrueFoundry doesn't reach that until its custom Enterprise tier (10M+, undisclosed pricing).
Even without any extra math, the implication is straightforward:
- FastRouter gives you a published, self-serve tier with a very high cap.
- TrueFoundry pushes high-volume use cases into custom Enterprise conversations earlier.
That matters operationally because “custom Enterprise” is usually where pricing stops being predictable, and where timelines slow down. If you expect growth, you should prefer a pricing curve that doesn't force you into a renegotiation the moment you succeed.
What FastRouter includes that isn't a line item (and why it matters at purchase time)
A lot of gateways play the same game: cheap entry tier, then the features you actually need in production (governance, guardrails, meaningful observability) are gated behind the expensive plans.
FastRouter's pricing is flat and token pass-through, and these come with it at every tier:
- Weekly cost insights — automated, evidence-backed savings recommendations surfaced from your own traffic, not a generic benchmark.
- Evaluation-verified model switching — routing/fallback changes are checked against your quality bar before they're suggested, not applied blindly.
- Governance on every tier — RBAC, project and key limits, and input/output guardrails aren't reserved for the top plan.
If those three things are real in the way you operate, they're not “nice-to-have.” They prevent gateway ownership from becoming a monthly grind:
- cost insights that are grounded in your traffic mean you're not manually spelunking dashboards trying to figure out why spend jumped;
- evaluation-verified switching is a guardrail against the classic failure mode where someone optimizes for cost and breaks output quality;
- governance on every tier means you don't have to buy your way into basic safety rails.
TrueFoundry does publish real-time cost observability — a metrics dashboard with per-token cost tracking, filterable by team, model, and user. What it doesn't publish is an automated, evidence-based weekly savings-recommendation engine, or an evaluation gate that checks a proposed model switch against quality before applying it. On guardrails specifically: TrueFoundry ships “built-in guardrails” on its free Developer tier, but full guardrail policies, RBAC, and audit logging are Pro/Enterprise features — not available on every tier the way they are on FastRouter.
How to decide (use this as a purchase framework)
1) Is self-hosted / VPC / on-prem / air-gapped a hard requirement?
- Yes → this doesn't eliminate either vendor. Both FastRouter and TrueFoundry offer a path here through their Enterprise tier — get a scoped quote from both rather than defaulting to TrueFoundry on the assumption it's the only option. TrueFoundry's self-hosted infra cost is published (~$600–$1,000/mo on top of platform fee); use that as your comparison baseline when evaluating FastRouter's Enterprise quote.
- No → keep going. If you can run managed, treat “managed + predictable cost” as the default because it removes an entire class of operational work.
2) Are you optimizing for lowest total platform overhead at production volume?
If your evaluation is “we're routing real traffic, what's the platform overhead per request,” the numbers above are decisive:
- FastRouter Pro: $0.00004 per request platform fee (plus zero token markup).
- TrueFoundry Pro: $0.0005 per request platform fee (about 12.5x higher per request than FastRouter Pro), before token costs.
If that's your primary buying criterion, FastRouter is the clear winner on published pricing.
3) Do you expect to scale beyond 1M requests/month without wanting a sales negotiation?
- If yes, watch the tier shapes: FastRouter gives you 5M at $199/mo and 50M at $799/mo as published tiers.
- TrueFoundry's published tiers top out at 1M requests/mo, and higher volume is custom (10M+ requests/mo) with undisclosed pricing.
- If you hate tier cliffs and procurement friction, prefer the product that publishes a high ceiling you can buy today.
4) Are you buying support/SLA, not throughput?
- If what you need is priority support and SLA, TrueFoundry makes that an explicit paid tier: Pro Plus at $2,999/mo with 1M requests/mo.
- If what you need is more headroom, Pro Plus doesn't solve that problem (same cap as Pro). Don't pay for the wrong thing.
5) Do you care about token pricing transparency?
- If you want clean, predictable token economics: FastRouter's “zero markup” and “provider's listed token price, always” is as straightforward as it gets.
- TrueFoundry does track cost per token internally — input and output tokens are costed separately, using continuously-updated provider rate cards — so the “request-based” framing describes what TrueFoundry charges you for platform access, not how granularly it measures the model spend passing through it. That model spend is visible in TrueFoundry's dashboard; it's TrueFoundry's own fee that's bucketed by request-tier rather than token count.
Deployment control doesn't decide this by itself — both vendors can meet an on-prem requirement through their Enterprise tier. What's left is “predictable subscription + zero token markup + cheap requests” (FastRouter) versus “request-tier capacity pricing with earlier tier cliffs” (TrueFoundry), and the published numbers above strongly favor FastRouter on cost at any real production volume.
Related Articles
.png&w=3840&q=75)
.png&w=3840&q=75)
Cheapest Way to Access GPT-6, Claude & Gemini Through One API
Find out about Cheapest Way to Access GPT-6, Claude & Gemini Through One API in new FastRouter blog

.png&w=3840&q=75)
.png&w=3840&q=75)
Stop Paying “Real Time” Prices for Work No One Needs in Real Time
See which of your traffic qualifies for Flex tier pricing, and exactly how much it saves. Same model, same output, lower bill.

.png&w=3840&q=75)
.png&w=3840&q=75)
From One SDK to a Cost Problem: Why Teams End Up Needing Routing Intelligence
Most teams do not start an AI roadmap with optimized inference spend at the top of the list. They start with a deadline.
