Portkey
portkey.ai
- Free plan
- Web
- API
- Self-hosted
What Portkey is
Portkey is a platform for managing, observing and governing generative AI across teams. Its gateway provides a unified API for accessing models and providers across modalities, with configurable routing, failovers, retries and load balancing. Monitoring tools include logs, traces, feedback, metadata, filters and alerts; prompt management adds templates, a playground, API endpoints, versioning and variables. Listed integrations include OpenAI, Anthropic, Google Gemini, Azure, Amazon Bedrock and Ollama, as well as agent frameworks such as CrewAI, LangChain, LangGraph and the OpenAI Agents SDK. The gateway offers simple and semantic caching for repeat requests. Security features include virtual provider keys, access controls, inbound IP and geographic rules, and sensitive-data redaction. Deployment options include self-hosted and local setups, while Enterprise lists private cloud and VPC hosting. The Developer plan is free and includes 10k recorded logs per month, with three days of log retention and 30 days of metrics retention. Production costs $49/month; Enterprise has custom pricing. Portkey lists its headquarters in San Francisco and launch in March 2023.
Who it suits
Portkey may suit teams building generative AI applications that need a shared model gateway, activity monitoring and prompt management. Self-hosting and Enterprise deployment options may also interest organizations with specific deployment or compliance needs.
Good to know, and what to watch
- Gateway supports routing, retries, failovers and load balancing.
- Logs, traces, alerts and filters support observability.
- Integrates with model providers and agent frameworks.
- Self-hosted and local deployment options are listed.
- Free-plan logs are retained for three days.
- Free plan includes 10k recorded logs monthly.
- Production overages cost $9 per additional 100k requests.
- Enterprise pricing is custom.
Plans
Every plan and priceEnterprise
Developer
Production
Free plan
Getting started
- Open the Portkey website
- Choose the free Developer plan or a paid plan
- Connect model providers such as OpenAI, Anthropic, Google Gemini, Azure, Amazon Bedrock or Ollama
- Configure gateway routing, retries, failovers or load balancing
- Create prompt templates and manage them through the playground or API
Limits to check
Developer includes 10k recorded logs per month, three days of log retention, 30 days of metrics retention and three prompt templates. Production includes 100k recorded logs monthly, with $9 charged for each additional 100k requests; Enterprise uses custom retention.
The verdict
Portkey combines AI gateway controls with observability, prompt tools and deployment choices.
Questions people ask
Is Portkey free?
Yes. The Developer plan is free forever and includes 10k recorded logs per month, three days of log retention, 30 days of metrics retention and three prompt templates.
How much does Production cost?
Production costs $49 a month and includes 100k recorded logs per month. Additional 100k requests cost $9.
What platforms can I use?
Portkey is offered through an API, as a self-hosted deployment and on the web.
Which providers and frameworks integrate?
Listed providers include OpenAI, Anthropic, Google Gemini, Azure, Amazon Bedrock and Ollama. Documented agent integrations include CrewAI, LangChain, LangGraph and the OpenAI Agents SDK.
Does Portkey offer self-hosting?
Yes. Self-hosted and local deployment options are listed. Enterprise also offers private cloud deployment and VPC hosting.
What support comes with each plan?
Developer lists community support, Production lists production support, and Enterprise lists dedicated onboarding and priority support.
What the maker publishes
- Agent ecosystem
- Documented integrations include agent frameworks such as CrewAI, LangChain, LangGraph, and the OpenAI Agents SDK. source
- Caching
- Portkey says its gateway offers simple and semantic caching to reduce latency and save costs on repeat requests. source
- Company
- Portkey’s company information page lists San Francisco, California as its headquarters and March 2023 as its launch date. source
- Compliance
- The Enterprise plan lists SOC 2 Type 2, GDPR, and HIPAA under advanced compliance, along with custom BAAs and data isolation. source
- Deployment
- Portkey lists self-hosted and local deployment options, and Enterprise offers private cloud deployment and VPC hosting. source
- Founded
- 2023 source
- Gateway
- Its AI gateway provides a unified API to access models and providers across modalities. source
- Guardrails
- The Production plan includes LLM and partner guardrails, while Enterprise lists custom guardrail hooks and advanced evaluation templates. source
- Headquarters
- San Francisco, California source
- Integrations
- The integrations catalog lists providers and platforms including OpenAI, Anthropic, Google Gemini, Azure, Amazon Bedrock, and Ollama. source
- Observability
- Portkey offers logs, traces, feedback, metadata, filters, and alerts for AI activity. source
- Product
- Portkey describes itself as a production platform for managing, observing, and governing GenAI across teams. source
- Prompt management
- Prompt management includes templates, a playground, API endpoints, versioning, and variables. source
- Routing
- The gateway supports configurable routing, failovers, retries, and load balancing across models. source
- Security
- Security features include virtual provider keys, access controls, inbound IP and geographic rules, and sensitive-data redaction. source
- Support
- The Free plan lists community support, Production lists production support, and Enterprise lists dedicated onboarding and priority support. source
Checked against the maker's pages on 29 Sept 2026.