Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

TrueFoundry AI Gateway

truefoundry.com

  • Free plan
  • Linux
  • Web
  • API
  • Self-hosted
8.0 Club score Visit truefoundry.com

What TrueFoundry AI Gateway is

TrueFoundry AI Gateway sits between applications and model providers to manage AI access, routing, governance, observability, security, and cost controls. One API supports chat, completion, embedding, and reranking models. TrueFoundry says the gateway connects to 1,000+ LLMs across 35+ platforms, including OpenAI, AWS Bedrock, Google Vertex AI, Azure OpenAI, and Anthropic. Routing options include latency-based selection, weighted load balancing, and automatic fallbacks to secondary models. The gateway tracks token use, latency, errors, and request volumes, with centrally stored request and response logs. Teams can apply rate limits and cost or token quotas, and manage use through role-based access control. MCP connections can link tools such as Slack, GitHub, Confluence, and Datadog, with OAuth2, RBAC, and metadata policies for tool calls. Safety guardrails include PII filtering and toxicity detection. The product is available as SaaS; Enterprise also lists VPC, on-premises, and air-gapped deployments. TrueFoundry says those deployment options can keep sensitive data within enterprise boundaries. Developer is free, Pro costs $25 a month per user, and Enterprise pricing is custom.

Who it suits

The gateway suits teams that need a shared API for model access, routing, usage visibility, and controls. Its deployment choices and access controls may also fit organizations with specific infrastructure or data-boundary requirements.

Good to know, and what to watch

  • One API supports four model types
  • Automatic fallback and weighted routing are available
  • Tracks usage, latency, errors, and request volumes
  • Free Developer plan is available
  • Self-hosted deployment options are listed for Enterprise
  • Developer allows up to 3 users
  • Developer includes 10,000 requests per user monthly
  • Pro charges for additional request usage
  • Enterprise pricing is custom

Enterprise

—

Developer

Free

Pro

$499/mo

Pro Plus

$2,999/mo

Free plan

Getting started

  1. Open the TrueFoundry AI Gateway website
  2. Choose SaaS or an available self-hosted deployment route
  3. Start with the free Developer plan or begin the free trial
  4. Connect model accounts and configure the unified API
  5. Set routing, quotas, access controls and guardrails

Limits to check

Developer includes up to 3 users, 10,000 requests per user per month, 5 model accounts and 5 MCP servers. Pro includes 20,000 requests per user per month and charges $20 per additional 100,000 requests; Enterprise uses custom usage terms.

The Club's review

TrueFoundry AI Gateway is a control layer for teams that need one managed route to many AI models. It suits developers and platform owners balancing provider choice with quotas, governance, and operational visibility. Its free Developer tier is useful for small deployments, while larger or security-sensitive organizations will need Pro or custom Enterprise terms.

Overview

TrueFoundry places an API gateway between applications and model providers, centralizing model access, routing, governance, observability, security, and cost controls. A unified API covers chat, completion, embedding, and reranking models, so applications can connect without handling each provider separately. TrueFoundry says it integrates more than 1,000 large language models across over 35 platforms, including OpenAI, AWS Bedrock, Google Vertex AI, Azure OpenAI, and Anthropic.

That breadth is most valuable when a team expects to change models or providers, or wants one policy layer across several services. A smaller project using one provider may not need the gateway's broader administration features.

Key features

Routing and resilience

Latency-based routing and weighted load balancing let teams distribute traffic deliberately. Automatic fallback to secondary models can keep requests moving when a preferred model is unavailable, making the gateway more suitable for production workloads than a single-provider integration.

Visibility and controls

Usage analytics cover token consumption, latency, errors, and request volume, with centrally stored request and response logs. Rate limits, cost or token quotas, role-based access control, virtual API keys, and budget controls give administrators practical ways to contain spend and separate usage by team or application.

Tools and safety

MCP integrations connect tools such as Slack, GitHub, Confluence, and Datadog. OAuth2, RBAC, and metadata policies govern those tool calls. Safety options include PII filtering and toxicity detection. Pro includes built-in guardrails; Enterprise adds built-in, custom, and third-party guardrails, which matters for organizations requiring tailored policies.

Pricing

Developer

Developer is free at $0 per user per month, with up to three users, 10,000 gateway requests per user per month, five model accounts, and five MCP servers. Support is through an in-app chatbot and Discord. This is a sensible starting point for a small team or evaluation, but the user, request, account, and server caps will constrain broader adoption.

Pro

Pro costs $25 per user per month, billed per user, with no user limit and 20,000 requests per user per month. Additional usage costs $20 per 100,000 requests. A support portal replaces the lighter Developer support channels. Pro fits growing teams that need more seats and capacity, though request overages add a variable cost that should be budgeted.

Enterprise

Enterprise uses custom pricing and custom usage terms. It adds advanced security, controls, and support, including priority support, an SLA, and a dedicated Slack channel. Enterprise deployment can run in a VPC, on premises, or in an air-gapped environment, helping organizations keep sensitive data within their own boundaries.

Platforms

TrueFoundry AI Gateway is offered through an API and web interface, with self-hosted deployment available. Enterprise customers can choose SaaS, VPC, on-premises, or air-gapped arrangements.

Who it's for

Choose it for a development or platform team standardizing access to many models while enforcing quotas, roles, logging, routing, and fallback behavior. Developer suits up to three users with modest traffic. Pro is the practical tier for organizations with unrestricted seats and higher per-user quotas. Enterprise is aimed at regulated or security-conscious deployments requiring private infrastructure and contractual support. It is less compelling for a small application that only needs a direct connection to one model provider.

Pros and cons

  • Pros: One API spans chat, completion, embedding, and reranking models across 1,000+ models and 35+ platforms.
  • Pros: Latency routing, weighted balancing, and automatic fallbacks improve control over availability and performance.
  • Pros: Detailed usage metrics, centralized logs, quotas, RBAC, and budget controls support accountable operation.
  • Pros: Enterprise private-deployment options and expanded guardrails address stricter security requirements.
  • Cons: Developer is limited to three users, 10,000 requests per user monthly, five model accounts, and five MCP servers.
  • Cons: Pro's $20 per additional 100,000 requests makes high-volume usage less predictable.
  • Cons: Enterprise pricing is custom, so costs cannot be assessed from the published tiers alone.

Alternatives

For a wider comparison, see LLM Gateway Software and Model Registry Software.

  • Bifrost is the better fit when you want an Apache 2.0, self-hosted open-source option deployable with npx, Docker, Kubernetes, or a Go binary.
  • Kong Gateway suits teams wanting a 30-day, no-credit-card trial with no gateway limits and 30 days of analytics retention.
  • LLM Gateway is an option for access to 200+ models across 40+ providers with a free tier charging 5% on credit usage.
  • Requesty fits lighter experimentation with 200 free requests per day, routing, caching, fallbacks, and EU data residency.
  • Helicone is a freemium, open-source alternative with a free Hobby tier.
  • LiteLLM is suited to self-hosted users seeking an open-source gateway with 100+ providers, virtual keys, budgets, and rate limits.
  • Vercel is a broader web platform rather than a dedicated AI gateway.
  • Envoy AI Gateway is a free, self-managed open-source project for Linux and macOS.

Verdict

TrueFoundry AI Gateway is a strong choice for teams that need multi-provider model access with routing, observability, quotas, and governance in one layer. Choose it when those controls justify Pro's per-user and usage costs or Enterprise's custom engagement. Look elsewhere if you only need a simple single-provider API or a fully open-source gateway without commercial-tier limits.

The verdict

TrueFoundry AI Gateway combines model connectivity with routing, monitoring, and access controls. The free Developer plan has user and request limits, while higher tiers add capacity and Enterprise deployment options.

Questions people ask

Is there a free plan?

Yes. Developer costs nothing and includes up to 3 users.

What does Pro cost?

Pro costs $25 a month, billed per user, with no user limit.

Which deployment options are available?

The gateway is available as SaaS and self-hosted. Enterprise also lists VPC, on-premises and air-gapped deployment.

Which model providers does it support?

TrueFoundry says it integrates with 1,000+ LLMs across 35+ platforms, including OpenAI, AWS Bedrock, Google Vertex AI, Azure OpenAI and Anthropic.

Does it support MCP tools?

Yes. MCP connections can include Slack, GitHub, Confluence and Datadog, with OAuth2, RBAC and metadata policies.

What support comes with each plan?

Developer lists an in-app chatbot and Discord, Pro lists a support portal, and Enterprise lists priority support, an SLA and a dedicated Slack channel.

What the maker publishes

Access controls
Teams can apply rate limits and cost or token quotas and use role-based access control to manage usage. source
API authentication
The metrics API accepts Personal Access Tokens or admin-created Virtual Account Tokens. source
API surface
The documentation exposes endpoints for chat, embeddings, reranking, responses, images, audio, batches, files, fine-tuning, moderation and model listing. source
Company
TrueFoundry lists its address as Ensemble Labs Inc, 355 Bryant Street, Suite 403, San Francisco, California 94107. source
Company size and location
LinkedIn lists TrueFoundry as privately held, headquartered in San Francisco, with 51–200 employees and a 2021 founding year. source
Configuration extensibility
Gateway policies can be managed in the UI or declared in YAML and applied through the TrueFoundry CLI for GitOps workflows. source
Cost controls
Usage visibility, token tracking, quotas, budgets, rate limits, caching and batching are provided to control inference spend. source
Deployment
The gateway supports SaaS and, for Enterprise, VPC, on-premises and air-gapped deployment options. source
Deployment options
The product is offered as SaaS and can also run in a customer VPC, on premises or in an air-gapped environment. source
Documentation depth
The docs include provider catalogs, authentication, self-hosted model setup, metrics APIs and a large endpoint reference covering multiple media and model operations. source
Founded
TrueFoundry states it was founded in 2021 by Nikunj Bajaj, Abhishek Choudhary and Anuraag Gutgutia. source
Funding
TrueFoundry announced a $19 million Series A led by Intel Capital, bringing reported total funding to $21 million. source
Guardrails
The product supports safety guardrails including PII filtering and toxicity detection; its pricing page lists built-in guardrails on Pro and built-in, custom and third-party guardrails on Enterprise. source
Guardrails
Configurable input and output controls can filter PII, detect prompt injection, assess toxicity and apply organization-specific safety policies. source
Headquarters
San Francisco, California, United States source
Limit: documentation and RBAC
G2 reviewers mention documentation lag or insufficient onboarding detail, missing edge-case examples and requests for stronger RBAC or deployment approvals. source
Limit: outage communication
Gartner reviewers repeatedly ask TrueFoundry to communicate its own downtime more proactively. source
Limit: pricing
G2 pricing analysis reports that some small businesses and startups consider the paid tiers expensive, especially for enterprise and on-premises deployments. source
MCP integrations
The gateway supports connecting tools such as Slack, GitHub, Confluence and Datadog, with OAuth2, RBAC and metadata policies for tool calls. source
MCP integrations
Native MCP support connects tools such as Slack, GitHub, Confluence and Datadog, with OAuth2, RBAC and metadata policies on tool calls. source
Microsoft Marketplace availability
TrueFoundry AI Gateway is listed in Microsoft Marketplace as an Azure product for governed, observable AI-agent deployment. source
Model coverage
TrueFoundry says the gateway connects more than 1,600 models, while its provider documentation lists integrations across more than 35 platforms and over 1,000 LLMs. source
Model providers
TrueFoundry says the gateway integrates with 1,000+ LLMs across 35+ platforms, including OpenAI, AWS Bedrock, Google Vertex AI, Azure OpenAI and Anthropic. source
Notable news
TrueFoundry announced TrueFailover in January 2026 to reroute enterprise AI traffic automatically during model-provider outages. source

Checked against the maker's pages on 15 Oct 2026.