Built by engineers who shipped Apigee. Production traffic on day one at HSBC, Fiserv, Zurich, MTN, Canara Bank and Mahindra Finance.
Gateways Were Built for a Different Decade
Whether you are on a heavyweight enterprise platform or a cloud bundle, the gateway you have today was designed for a problem shape that no longer matches what most teams run.
API Estates Exploded
The average large enterprise now runs hundreds of APIs across multiple business units, gateways and clouds. Existing platforms were designed when an org might run a dozen.
The Consumers Changed
APIs used to be called by apps you controlled. Now they are called by partners, internal teams, and a fast-growing population of AI agents.
The Deployment Surface Shifted
Edge, multi-region, sovereign cloud, hybrid. Yesterday's gateway runs in one VPC behind one load balancer. That is no longer the shape most teams need.
The Cloud Bundle Stopped Being Enough
AWS API Gateway, Azure APIM and Cloudflare gave you serverless routing, until you needed a portal, partner onboarding, monetization or governance. None of that was on the roadmap.
Scale
Built for the Traffic You Are Actually Seeing
Predictable latency under load, linear horizontal scale, stateless by design. The numbers below are reproducible because Helix is written in Rust and the request path does not allocate on the heap.
Latency
<1ms
p99 under load
At 50,000 RPS per core. Predictable tail latency, not a best-case demo number.
Throughput
Millions RPS
The consumers changed
Stateless nodes. No shared cache, no warm-up. Add capacity, get capacity.
Uptime
99.99%
Multi-region active-active
Designed to lose a region without losing the request in flight. Real SLA.
Footprint
20 MB
Single static binary
Runs in 32 MB of memory. Deploys to Kubernetes, edge runtimes, or a bare VM.
.avif)
.png)
Helix is built in Rust, compiled to native code. That is why the numbers above are reproducible, not best-case marketing.
Everything a Gateway Should Do!
The foundation a serious gateway is expected to ship with. Routing, security, lifecycle and observability, built in, not bolted on through plugins or third-party consoles.
Network
Global Protection, Layer by Layer
Data protection and compliance for clusters in every region you operate, applied per layer, not retrofitted at the edge.
Lifecycle
The Whole Lifecycle, One Surface
Design, document, mock, govern, version and deprecate. Gateway, portal and docs, operated from a single control plane.
Clusters
One Dashboard, Every Cluster
Manage clusters across regions, environments and clouds from one console. No region-by-region tab-switching.
Identity
Auth That Plugs Into What You Run
Built-in RBAC plus first-class support for OAuth, OIDC, Keycloak, Okta, 2FA and mTLS. No proprietary identity layer.
Observability
Telemetry Where You Already Look
OpenTelemetry-first. Native exporters for Prometheus, Grafana, Datadog and structured JSON audit logs.
Deploy
Run It Where You Run
Bare metal, virtual machines, Kubernetes, edge or cloud. Same binary, same behaviour, no rewrite per surface.
Docs
Docs the Moment You Publish
Auto-generate API documentation from your spec and publish it to a portal third-party developers can actually consume.
Legacy
Legacy Soap, Fronted as Rest
Translate modern RESTful calls to the original SOAP service untouched. No backend changes, no rewrite required.
Routing
Tag Traffic Before You Route It
Classify by consumer, intent or risk. Then route, throttle or audit using the same label downstream.
And Everything That Should Come Next!
Most gateways stop at the foundation above. Helix keeps going, a self-branded developer portal, monetization that actually bills, and MCP endpoints for agents, all from the same published API.
Self-branded portal
A Developer Portal, Wired to the Gateway
Publish an API to Helix and it shows up in your branded developer portal automatically. Sign-up, app creation, key issuance, sandbox access and live usage dashboards. The portal is not a separate platform you integrate, it is the gateway, surfaced to your developers.
%20(1).png)
Monetization
Plans, Metering and Billing, Included
Define tiers, attach quotas, track usage per consumer, push invoicing data into your billing system. From free trial to paid partner contract, no second platform.
Agent-ready
MCP Endpoints, the Day You Publish
Every API becomes available as an MCP tool with scoped agent credentials, separate rate limits, and audit trails for non-human traffic.
%20(1).png)
Who Helix is for
Three Teams Helix Is Built For
Different starting points, same destination. Helix fits each without forcing a rewrite or a long migration.
Replace
Manual Onboarding Doesn’t Scale
Apigee, MuleSoft, IBM API Connect. Outgrew the cost, the operational complexity, or the velocity. Helix replaces them with a lighter, faster foundation in 3 to 4 days policy mapping handled by our team. No rewriting APIs, no downtime.
Escape
Teams Locked Into Proprietary Config
If your gateway has its own policy language, deployment model and runtime, you are not really portable. Helix is OpenAPI-native and vendor-neutral by design.Your policies, your specs, your data yours.
Greenfield
Teams Choosing Their First Serious Gateway
On AWS API Gateway, Azure APIM or Cloudflare today. Hit the limits the bundle was never going to cover portal, monetization, partner onboarding, agents. Skip a generation. Start with the architecture you would otherwise inherit and have to replace.
Self-managed
Run by Your Engineers.
Not a Certification Program
Existing gateways assume you will hire specialists. Lua engineers for Kong. Apigee-certified consultants. MuleSoft architects. That assumption belongs to a different decade.
Helix is built so the engineers you already have can run it. No proprietary policy language. No plugin compilation toolchain. No three-week training course before someone can ship a route.
In an AI-native operating model, the gateway should be operable by a small team working with AI tooling, not a dedicated platform group with niche skills. Helix is designed for exactly that shape of team.
You should not need to grow the team to grow the gateway.The gateway should grow with the team you have.
No custom plugin
Get in-built plugin compilation, no custom DSL, no scripting language to maintain.
No certifications
No vendor-issued credentials required to ship a route or write a policy.
No specialists
No dedicated platform team. Your backend engineers operate it.
Typical team to operate
1 – 2
Engineers on Helix, working alongside AI tooling for routine ops. Compared to 4 to 8 dedicated platform engineers on a heavyweight gateway.
Budget impact
$600k – $1.4M / yr
In fully-loaded platform headcount saved, before you count the licence fee difference.
With DigitalAPI, Implementing Fintech Onboarding and API monetization has been a game-changer for our business.
%20(1).avif)
Sovdeep Das
Sr. Director, Product Management, Fiserv
How Helix Compares?
A practical read of the four gateways most teams are choosing between. We have stayed on the dimensions that actually decide the bake-off.
Dimension
.avif)
Modern
Heavyweight
Open-core
Cloud bundle
Pricing per 100k transactions
Indicative
$2,000
Enterprise (opaque)
$8,000 – 10,000
Bundled with AWS
p99 latency
Under sustained load
< 1 ms
~10 ms
~5 ms
~25 ms
Specialists required
To operate at scale
No specialists required
Apigee-certified consultants
Lua engineers
AWS specialists
Self-branded developer portal
Built in, branded, self-serve
.svg)
API monetization
Plans · metering · billing
AI / MCP native
Agents as first-class consumers
Multi-cloud
Beyond a single provider
Vendor lock-in
Proprietary config & runtime
High (AWS-shaped)
High
Medium
High (AWS-shaped)
Time to migrate
From similar gateway
Less than
1 week*
—
—
—
Pricing per 100k transactions
Indicative
Modern
$2,000
Heavyweight
Enterprise (opaque)
Open-core
$8,000 – 10,000
Cloud bundle
Bundled with AWS
p99 latency
Under sustained load
Modern
< 1 ms
Heavyweight
~10 ms
Open-core
~5 ms
Cloud bundle
$8,000 – 10,000
Specialists required
To operate at scale
Modern
AWS specialists
Heavyweight
Apigee-certified consultants
Open-core
Lua engineers
Cloud bundle
AWS specialists
Self-branded developer portal
Built in, branded, self-serve
Modern
Heavyweight
Open-core
Cloud bundle
API monetization
Plans · metering · billing
Modern
Heavyweight
Open-core
Cloud bundle
AI / MCP native
Agents as first-class consumers
Modern
Heavyweight
Open-core
Cloud bundle
Multi-cloud
Beyond a single provider
Modern
Heavyweight
Open-core
Cloud bundle
Vendor lock-in
Proprietary config & runtime
Modern
High (AWS-shaped)
Heavyweight
High
Open-core
Medium
Cloud bundle
High (AWS-shaped)
Time to migrate
From similar gateway
Modern
3-4 days
Heavyweight
—
Open-core
—
Cloud bundle
—
.avif)
The team
The Engineers Who Built Apigee, Building What Comes Next
Helix is built by a team that has shipped this category before. The credibility is earned, not asserted.
%20(1).png)
Apigee · 2004 → 2016
Apigee's first engineer. Built the original gateway core. Now building it again, lighter, faster, more open.
Apigee · 2007 → 2018
Led Apigee's runtime team through the Google acquisition. Authored the policy engine many of you still use.
Apigee · 2009 → 2020
Shipped Apigee's portal product. Knows where the developer-experience gaps are, because he lived inside them.
Apigee · 2011 → 2019
Ran Apigee's monetization product. Helix's pricing and metering layer is hers, built right this time.
Live portals
From Apigee, Kong or AWS in Less Than a Week, at Zero Cost
We import your existing config, run shadow traffic, hand you the diff, and cut over only when you say. Every migration above is real and signed off by a real customer.
Financial Services
< 1 week
From spec import to production cutover.
We absorb the engineering effort.
Migration cost
$0
No professional services line item. No "implementation partner." We do the work alongside your team.
Rollback path
30 days
Your old gateway stays live alongside Helix until you sign off. Then we wind it down on your timeline.
From Apigee
Classic · X · Edge
Import proxies, policies, environments and key value maps automatically.
Shadow traffic, diff response payloads, fix the deltas before cutover.
Cut DNS over a weekend. Apigee stays live in standby for 30 days.
We moved 47 APIs in a weekend. Monday morning, the only people who noticed were in finance.
From Kong
OSS · Konnect · Enterprise
Import services, routes, consumers and plugins (where 1:1).
Rebuild any custom Lua plugins on our cloud-native runtime.
Shadow + diff. Cut over on day 4 with full rollback path.
No more custom plugins to maintain. Our backend team operates the gateway now, without a specialist.
From AWS API Gateway
REST · HTTP · WebSocket
Import REST/HTTP APIs, stages, authorizers, usage plans.
Wire portal, monetization and MCP — none of which AWS shipped.
Route 53 cutover with weighted traffic. Roll back any time.
We were on AWS API Gateway for years. The portal alone was worth the move.
.avif)
