PortKey
LLM Gateways & APIs · Free + paid
AI gateway with observability and fallback routing
At a glance
- Made by
- Portkey (now offered as Prisma AIRS AI Gateway)
- Korean
- None (English and a few others only)
- API
- Yes
- Open source
- An open-source version is available (the license name is not on the page)
What is PortKey?
Portkey provides the AI gateway, observability, guardrails, governance, and prompt management that AI teams need to go to production, all in one platform. It gives access to more than 1,600 LLMs through a unified API and supports fallbacks, load balancing, and retries. Integration takes just a few lines of code, and every LLM request is monitored. It states that it is now generally available to enterprises as 'PRISMA AIRS AI Gateway'.
- Access 1,600+ LLMs through a unified API
- Automatic fallbacks, load balancing, and retries
- Log and trace observability dashboard
- Guardrails (deterministic, LLM, partner)
- Prompt templates, version control, playground
- Simple and semantic caching
- Governance including RBAC and SSO
- Open-source self-hosted version
Who uses PortKey and for what
- Teams deploying LLM apps to production
- Large-scale enterprise workloads that require compliance
- Evaluating prototypes and enterprise POCs
Pricing
What you get for free — Yes (Developer is free forever, 10k recorded logs per month, logs kept for 3 days)
| Plan | Price | Main limits |
|---|---|---|
| Developer | $0/mo | 10k recorded logs per month, logs kept 3 days and metrics 30 days, 3 templates |
| Production | $49/mo | 100k recorded logs per month, +$9 per 100k requests (up to 3M), logs kept 30 days |
| Enterprise | Contact sales | Custom pricing, 10M+ logs per month, VPC·private cloud |
| Open Source | 0 | Self-hosted, basic dashboard, community support |
Checked on the official pages on 2026-10-14. Prices and features can change, so check the official site again before paying. · Official pricing page
Combos that work well with PortKey
Per-model cost, quality, and incident response
ChatGPT routes between low-cost and high-performance models based on task difficulty and expected tokens. In the LiteLLM + OpenRouter + PortKey combo, the gateway consistently applies key management, call logs, and fallback model policies. Sample evaluations, error rates, and costs are compared to adjust routing rules periodically.
Cost: API pay-as-you-go + subscription (ChatGPT) / open source + mixed pay-as-you-go/subscription
Per-model cost, quality, and incident response
Claude routes between low-cost and high-performance models based on task difficulty and expected tokens. In the LiteLLM + OpenRouter + PortKey combo, the gateway consistently applies key management, call logs, and fallback model policies. Routing rules are adjusted periodically by comparing sample evaluations, error rates, and cost.
Cost: Freemium (free + paid subscription) + API pay-as-you-go / open source + mixed pay-as-you-go/subscription
Per-model cost, quality, and incident response
Gemini routes between low-cost and high-performance models based on task difficulty and expected tokens. In the LiteLLM + OpenRouter + PortKey combo, the gateway applies key management, call logs, and fallback model policies consistently. Routing rules are adjusted periodically by comparing sample evaluations, error rates, and costs.
Cost: Freemium (free + paid subscription) + API pay-as-you-go / open source + mixed pay-as-you-go/subscription
Per-model cost, quality, and incident response
DeepSeek routes between low-cost and high-performance models based on task difficulty and expected tokens. In the LiteLLM + OpenRouter + PortKey combo, the gateway applies key management, call logs, and fallback model policies consistently. Routing rules are adjusted periodically by comparing sample evaluations, error rates, and costs.
Cost: Open source (free) + API pay-as-you-go / open source + mixed pay-as-you-go/subscription
Per-model cost, quality, and incident response
Perplexity routes between low-cost and high-performance models based on task difficulty and expected tokens. In the LiteLLM + OpenRouter + PortKey combo, the gateway applies key management, call logs, and fallback model policies consistently. Sample evaluations, error rates, and costs are compared to adjust the routing rules periodically.
Cost: Freemium (free + paid subscription) / open source + mixed pay-as-you-go/subscription
Per-model cost, quality, and incident response
Grok-2 routes between low-cost and high-performance models based on task difficulty and expected tokens. The LiteLLM + OpenRouter + PortKey combo has the gateway apply key management, call logs, and fallback model policies consistently. Sample evaluations, error rates, and costs are compared to adjust routing rules periodically.
Cost: Subscription (including X subscription) + API pay-as-you-go / Open source + mixed pay-as-you-go/subscription
Per-model cost, quality, and incident response
Llama routes tasks to low-cost or high-performance models based on task difficulty and expected tokens. In the LiteLLM + OpenRouter + PortKey combo, the gateway consistently applies key management, call logging, and fallback model policies. Routing rules are adjusted periodically by comparing sample evaluations, error rates, and cost.
Cost: Open source (free, license conditions apply) / Open source + mixed pay-as-you-go/subscription
Per-model cost, quality, and incident response
Mistral routes tasks to low-cost or high-performance models based on task difficulty and expected tokens. In the LiteLLM + OpenRouter + PortKey combo, the gateway consistently applies key management, call logging, and fallback model policies. Routing rules are adjusted periodically by comparing sample evaluations, error rates, and cost.
Cost: Open source (free) + API pay-as-you-go / open source + mixed pay-as-you-go/subscription
AI tools similar to PortKey (alternatives)
PortKey vs. alternatives at a glance
| AI | Pricing | Combos on the map | Connected tools | Free | Cheapest monthly | Korean |
|---|---|---|---|---|---|---|
| PortKey | Free + paid | 10 | 5 | Yes | $49/mo | No |
| Cloudflare AI Gateway | Free + paid | 0 | 5 | Yes | — | Yes |
| LiteLLM | Free | 10 | 5 | No | — | No |
| OpenRouter | Free + paid | 10 | 5 | Yes | — | No |
| Vercel AI Gateway | Free + paid | 10 | 5 | Yes | $5/mo | No |
4 of these 5 tools have a free plan. Cheapest paid plan: Vercel AI Gateway ($5/mo).
Combo and connection counts are based on the relationships recorded on AI Galaxy Map (2026-10-26); they are not a quality ranking.
PortKey on AI Galaxy Map
- One of 14 AI tools in the LLM Gateways & APIs field.
- Connected to 5 other tools — #6 in its field.
- Part of 10 work combos recorded on the map.
Frequently asked questions
Can I use PortKey for free?
Yes (Developer is free forever, 10k recorded logs per month, logs kept for 3 days). Prices can change, so check the official site before paying.
What works well with PortKey?
Based on the combos on the map, there is a workflow such as "Per-model cost, quality, and incident response" that uses it together with GPT, LiteLLM, OpenRouter.
What AI can I use instead of PortKey?
Tools that do the same job, such as Cloudflare AI Gateway, LiteLLM, OpenRouter, Vercel AI Gateway, are connected to it on the map.
Can I use PortKey in Korean?
None (English and a few others only)
Sources (official pages)
Checked on the official pages on 2026-10-14. Prices and features can change, so check the official site again before paying.