OrcaRouter
OrcaRouter is an AI gateway, operated by CONTINUUM AI PTE. LTD., that routes requests across 200+ LLMs via an OpenAI-compatible API. It serves AI application developers, AI-native product teams, and regulated enterprises through a freemium model with zero token markup.
- Company typePrivate
- Founded2026
- HeadquartersSan Francisco, United States
- Headcount—
- GTM typeB2B
- OfferingSoftware
What OrcaRouter does
OrcaRouter is an AI gateway platform operated by CONTINUUM AI PTE. LTD., a Singapore-incorporated private limited company, that exposes an OpenAI-compatible API for routing large language model (LLM) requests across more than 200 underlying models. The platform abstracts multi-model access behind a single endpoint and integrates directly with widely used AI development tools including Cursor, Cline, Continue, Aider, Vercel AI SDK, LangChain, and LlamaIndex. Its proprietary routing layer selects models based on context, cost, and performance objectives using contextual embeddings and online learning from real traffic; routing accuracy is published on the RouterArena leaderboard at 75.5% as of June 2026.
The product surface comprises the core routing platform, the Fusion family (Fusion, Fusion Mini, Fusion Flash), an MCP Server, Agent Firewall and Guardrails, a Routing DSL, the public RouterArena leaderboard, a Playground, and a Rewards Program. An open-source distribution, OrcaRouter-Lite, is available on GitHub. The platform holds GDPR, SOC 2 Type II, HIPAA, and ISO 27001 certifications, with cloud infrastructure hosted in Singapore and the United States. No publicly disclosed funding rounds, headcount figures, revenue figures, or named management executives appear in the source material.
The business model is a zero-token-markup, freemium structure: a free "Hacker" tier, a "Team" tier at $499 per month, and custom Enterprise contracts. The company serves AI application developers, AI-native product teams, and enterprise customers with regulated-data requirements. Revenue mechanics are anchored on subscription pricing rather than token margin, which differentiates the unit economics from typical gateway competitors that apply 5-20% markups on token pass-through.
OrcaRouter firmographics
Firmographics- Name
- OrcaRouter
- Legal name
- CONTINUUM AI PTE. LTD.
- Website
- https://orcarouter.ai
- Company type
- Private
- Founded year
- 2026
- Operating status
- Operating
- Short description
- OrcaRouter is an AI gateway, operated by CONTINUUM AI PTE. LTD., that routes requests across 200+ LLMs via an OpenAI-compatible API. It serves AI application developers, AI-native product teams, and regulated enterprises through a freemium model with zero token markup.
- Ownership category
- akta.pro rank
Where OrcaRouter is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
OrcaRouter business model
Business model- GTM type
- B2B
- Offering type
- Software
Revenue model
- Token Pass-through (Zero Markup): Customers pay upstream LLM providers directly at their published rates. OrcaRouter adds $0 per token. This is the core product value proposition — not a revenue stream but a competitive differentiator that drives adoption.
- Hacker Tier (Free): Full gateway access: 200+ models, one OpenAI-compatible API, automatic failover, basic observability, prompt versioning, 3 API keys. Free forever with zero token markup.
- Team Subscription: $499/month. Everything in Hacker plus: up to 10 team seats, compliance enforcement and reports, unlimited API keys, priority support. New workspaces get first month free.
- Enterprise Subscription: Custom pricing. Private/on-prem deployment, 99.99% uptime SLA, dedicated infrastructure, dedicated support. Volume-based pricing; contact sales. May include flat per-million-token routing fee depending on volume tier.
- Rewards Program (Points): Points accumulate from API calls (1 point per 1,000 tokens). Points never expire. Tiers: Ripple (free), Wave (1K pts - priority routing), Surge (10K pts - 1.5x rate limits + analytics), Tide (50K pts - 2x rate limits + dedicated support), Storm (200K pts - 3x rate limits + revenue share from platform margin). 7-day streak multiplier: 1.25x points.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Pay-as-you-go | Free tier with full gateway access and zero token markup |
| Subscription | Monthly | Team collaboration features for growing teams |
| Subscription | Multi-year contract | Enterprise-grade deployment with SLA commitments |
Go-to-market motion2 records
Distribution channels2 records
Marketing channels4 records
OrcaRouter product offering
Product offeringCore offering
OrcaRouter is an OpenAI-compatible AI gateway that exposes a single API endpoint to access 200+ large language models from providers such as OpenAI, Anthropic, Google, DeepSeek, Together AI, and Groq. It grades every prompt using contextual embeddings and online learning to route each request to the optimal model, and provides automatic mid-stream failover, load balancing, guardrails, an agent firewall, prompt versioning, caching, and per-request observability. Customers pay each upstream provider's published token rates directly with zero markup, with optional Team and Enterprise subscriptions for collaboration, compliance, and SLA features.
Product overview
OrcaRouter is an AI gateway platform built by CONTINUUM AI PTE. LTD. (Singapore) that provides a single OpenAI-compatible API endpoint routing across 200+ LLM models from OpenAI, Anthropic, Google, DeepSeek, Alibaba Cloud, Zhipu AI, Moonshot, Kuaishou (Kling), and others. The core platform (OrcaRouter) combines adaptive prompt routing, automatic failover, guardrails, and an agent firewall. Security products include Guardrails (PII Shield, injection detection) and the Agent Firewall (six-verdict tool/MCP gating, anomaly detection) — both made free for all users in June 2026. The Fusion sub-product family (Fusion, Fusion Mini, Fusion Flash) runs multiple models in parallel with a judge LLM to select or synthesize outputs at Fable-5-level quality. The Routing DSL enables programmable inference graphs. Ancillary products include the RouterArena public leaderboard, the Playground for model testing, and the Rewards Program loyalty tier. An open-source version (OrcaRouter-Lite) is available on GitHub.
Differentiator
Problem solved
Functional benefit
Products and services
- OrcaRouter Core Platform An OpenAI-compatible AI gateway that routes LLM requests across 200+ models from OpenAI, Anthropic, Google, DeepSeek, and others through a single endpoint. Provides adaptive routing, automatic failover, load balancing, guardrails, an agent firewall, observability, prompt versioning, and caching at zero token markup for AI engineering teams building production applications.
- OrcaRouter MCP Server A native Model Context Protocol server that lets AI agents and tools route requests through OrcaRouter via MCP dispatch rather than a direct SDK integration, with tool-gating, audit, and anomaly detection at the MCP layer.
- OrcaRouter Agent Firewall Action-plane security control for AI agents that grades every tool call, MCP dispatch, and network egress with six verdicts (ALLOW, REVIEW, BLOCK, sanitize, pending-approval, cap-cost) before execution, with scoped per-agent identity, MCP gating, and anomaly detection. Made free for all users.
- OrcaRouter Guardrails Content-plane security controls enforced inline before upstream calls are billed. Includes PII Shield, injection and jailbreak detection via an LLM judge, secret blocking, and output screening; blocked requests return a clean 400 and are never charged. Made free for all users.
- OrcaRouter Fusion Curated multi-model router that runs Claude Opus 4.8, GPT-5.5, and Gemini 3.1 Pro in parallel on hard prompts and uses Claude Opus 4.8 as a judge to select the strongest answer, returning it verbatim. Targets Fable-5-level quality at a fraction of the cost with a 1M-token context window.
- OrcaRouter Fusion Mini Two-model Fusion panel combining Claude Opus 4.8 and GPT-5.5 with Claude Opus 4.8 as judge, targeting balanced Fable-5-level inference at lower cost than the full Fusion tier with a 1M-token context window.
- OrcaRouter Fusion Flash Budget-tier Fusion panel combining Gemini 3.5 Flash, MiniMax M2.7, and GLM 5.1 with Claude Opus 4.8 as judge, targeting fast and cheap Fable-5-level inference for high-volume workloads with a 200K-token context window.
- OrcaRouter Rewards Program Points-based loyalty program where users earn 1 Point per 1,000 tokens routed (1.25x multiplier on a 7-day streak). Five tiers (Ripple, Wave, Surge, Tide, Storm) provide higher rate limits, priority routing, advanced analytics, dedicated support, and a planned revenue-share pool for Storm members. Points never expire.
Quantifiable outcome
- Up to 40% lower AI inference costs
- +4 more outcomes
Companies that use OrcaRouter
Customer profileSegments3 records
Ideal customer profiles3 records
OrcaRouter technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration16 records
AI capability13 records
Feature11 records
OrcaRouter partnerships and signals
Strategic signalPartnerships
15 partnerships are on record, tiered core.
- CursorcoreAI-first code editor integration. Setup guide provided. Developers can route Cursor through OrcaRouter to access all 200+ models with automatic failover.
- ClinecoreAutonomous coding agent for VS Code. Integration guide available. OrcaRouter provides unified access to multiple models for Cline's agentic coding workflows.
- ContinuecoreOpen-source AI code assistant for VS Code and JetBrains. Setup guide available. Enables developers to use OrcaRouter routing from their preferred IDE.
- AidercoreAI pair programming in terminal. Integration guide available. OrcaRouter routes Aider requests to optimal models based on task complexity.
- Vercel AI SDKcoreStreaming AI SDK for TypeScript applications. Setup guide available. Drop-in OrcaRouter endpoint for Vercel-deployed AI apps.
- LangChaincoreCompose LLM apps in Python or JavaScript. Integration guide available. OrcaRouter provides routing layer across multiple models within LangChain workflows.
- LlamaIndexcoreData framework for LLM applications. Setup guide available. OrcaRouter routes RAG and query workloads to optimal models.
- OpenAI SDKcoreDrop-in OpenAI-compatible API. Set base_url to api.orcarouter.ai/v1 and swap API key. All existing code works unchanged.
- Google GenAI SDKcoreGoogle's official GenAI SDK integration. OrcaRouter routes requests to Gemini and other models alongside OpenAI-compatible endpoints.
- Anthropic SDKcoreAnthropic's official SDK integration. OrcaRouter routes Claude requests alongside other providers with zero markup.
- OpenAIcoreUpstream LLM provider routed to via OrcaRouter. Customers pay OpenAI's published rates directly with zero markup.
- AnthropiccoreUpstream LLM provider routed to via OrcaRouter. Anthropic Direct routing available. Customers pay Anthropic's published rates directly with zero markup.
- GooglecoreUpstream LLM provider (Gemini) routed to via OrcaRouter. Google Direct routing available. Customers pay Google's published rates directly with zero markup.
- Together AIcoreUpstream LLM provider routed to via OrcaRouter. OpenRouter compatible ecosystem partner.
- GroqcoreUpstream LLM provider routed to via OrcaRouter. Fast inference provider integrated for low-latency routing.
Scale indicators7 records
Recent moves5 records
Expansion highlights6 records
OrcaRouter competitors and assessment
Company assessmentBroad incumbents
- Microsoft Azure: Azure AI Foundry and Azure OpenAI Service provide multi-model access with routing across OpenAI and partner models, and already score 72.8% on RouterArena. Broad incumbent competing on multi-model gateway capabilities.
- Amazon Web Services: AWS Bedrock provides multi-model foundation model access with unified routing and governance as part of its cloud platform. Broader incumbent offering overlapping gateway functionality bundled with cloud spend.
- Cloudflare: Operates Cloudflare AI Gateway as part of its broader edge and security portfolio, offering LLM routing, caching, and analytics. A larger incumbent with overlapping gateway capabilities but no specialization in deep adaptive routing.
- Google Cloud: Vertex AI offers multi-model access (Gemini, Anthropic, open models) with routing, governance, and enterprise controls. Broad incumbent that also serves as an upstream provider routed through OrcaRouter.
Direct peers
- Helicone: LLM observability and routing platform for AI applications, offering request logging, caching, and model selection. Overlaps with OrcaRouter on routing and per-request observability for production LLM apps.
- LiteLLM: Open-source LLM gateway providing unified API access, routing, load balancing, and failover across 100+ models. Comparable self-serve, developer-first product with overlapping routing and governance features.
- Portkey: AI gateway with LLM routing, observability, failover, and guardrails targeting production AI applications. Competes head-to-head with OrcaRouter on routing intelligence and enterprise governance features.
- OpenRouter: Multi-model LLM routing gateway offering a single API endpoint across hundreds of models from OpenAI, Anthropic, Google, and others. Closest direct competitor to OrcaRouter on routing, pricing structure, and developer onboarding.
- NotDiamond: LLM model routing decision platform that selects the optimal model per request. Direct competitor in the routing accuracy leaderboard space where OrcaRouter currently leads.
- Martian: LLM routing optimization company whose router is tracked on RouterArena (61.6% accuracy). Direct competitor in adaptive model routing for multi-model AI applications.
Market position
Strengths5 records
Weaknesses4 records
Competitive moat8 records
Key risks5 records
Key highlights6 records
Customer concentration
OrcaRouter social profiles
Digital presenceOrcaRouter compliance and trust
Trust signalCompliance4 records
OrcaRouter financial estimates
Financial estimateRevenue estimate
Valuation estimate
OrcaRouter leadership team
Management profileNumber of profiles
OrcaRouter funding detail
Funding detailFunding overview
Funding rounds
Investors
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
OrcaRouter M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about OrcaRouter
What does OrcaRouter do?
OrcaRouter is an OpenAI-compatible AI gateway that exposes a single API endpoint to access 200+ large language models from providers such as OpenAI, Anthropic, Google, DeepSeek, Together AI, and Groq. It grades every prompt using contextual embeddings and online learning to route each request to the optimal model, and provides automatic mid-stream failover, load balancing, guardrails, an agent firewall, prompt versioning, caching, and per-request observability. Customers pay each upstream provider's published token rates directly with zero markup, with optional Team and Enterprise subscriptions for collaboration, compliance, and SLA features.
Is OrcaRouter a public or private company?
OrcaRouter is a private company. It is classified as founder individual operated bootstrapped and is currently operating.
When was OrcaRouter founded?
OrcaRouter was founded in 2026.
Where is OrcaRouter based?
OrcaRouter is headquartered in San Francisco, United States, in the North America region.
How does OrcaRouter make money?
Five revenue lines are on record. Token Pass-through (Zero Markup) is the primary driver. The others are hacker Tier (Free), team Subscription, enterprise Subscription and rewards Program (Points).
Who are OrcaRouter's main competitors?
Broad incumbents on record are Microsoft Azure, Amazon Web Services, Cloudflare and Google Cloud. Direct peers are Helicone, LiteLLM, Portkey, OpenRouter, NotDiamond and Martian.
Does OrcaRouter have an API?
Yes. OrcaRouter exposes a single OpenAI-compatible API endpoint (https://api.orcarouter.ai/v1) that serves as an AI gateway across 200+ models. Developers point their existing OpenAI SDK at the OrcaRouter base URL and pass an OrcaRouter API key; no other code changes are required. The endpoint supports adaptive routing, load balancing, automatic mid-stream failover, prompt versioning, caching, guardrails, agent firewall, and per-request observability logs. No per-token markup is added — users pay each upstream provider's published rate directly. Python, TypeScript, and cURL code samples are provided on the product pages. Developer documentation is at docs.orcarouter.ai.