Portkey
"Production Stack for Gen AI Builders" [1]
Portkey is a production infrastructure layer for generative AI teams, providing a unified REST API across 1,600+ models with built-in routing, automatic fallback, load balancing, semantic caching, and observability logging. It targets developers and enterprises building LLM-powered applications who need cost controls, prompt versioning, and AI guardrails including PII redaction. Paid plans start at $49 per month with a free tier capped at 10,000 logged requests; an open-source self-host option is available under the MIT license. Portkey holds SOC 2 Type II, HIPAA, GDPR, and ISO 27001 certifications, though compliance certificates and private VPC deployments are restricted to the Enterprise tier.
Best for / Avoid if
Best for: Prototypes and side projects - free to start, no sales call; Regulated or enterprise workloads - compliance attestations and an enterprise plan; AI agents and automation - an agent-ready surface (MCP / llms.txt)
Pricing & procurement
- Pricing model
- Hybrid (base + usage) [2]
- Published pricing
- ✓ Yes [3]
- Free tier
- ✓ Yes [4]
- Free tier details
- Developer plan: free forever, 10k recorded logs/month, 3-day log retention, 30-day metrics retention, community support. Also open-source self-hosted gateway available (MIT license, unlimited requests, basic dashboard). [5]
- Self-serve signup
- ✓ Yes
- Requires sales call
- ✗ No
- Enterprise plan
- ✓ Yes [6]
| Plan | Item | Per | Amount | Source |
|---|---|---|---|---|
| Open Source | Self-hosted gateway (open source, MIT license) | $0 | source | |
| Developer | Platform plan fee | month | $0 | source |
| Developer | Recorded logs included | 10,000 logs/month | $0 | source |
| Production | Platform plan fee | month | $49 | source |
| Production | Recorded logs included | 100,000 logs/month | $0 | source |
| Production | Log overage | 100,000 additional requests | $9 | source |
| Enterprise | Platform plan fee | month (custom pricing) | - | source |
Capabilities
- Supported actions
- unified_chat_completions, openai_compatible_api, model_routing, automatic_fallback, load_balancing, semantic_caching, simple_caching, prompt_caching, spend_limits, budgets, rate_limiting, observability_logging, tracing, guardrails, pii_redaction, virtual_keys, byo_provider_keys, prompt_management, prompt_versioning, conditional_routing, canary_testing, automatic_retries, circuit_breaker, request_timeout, multimodal_support, embeddings_routing, image_generation_routing, audio_routing, rerank_routing, realtime_api_websocket, grpc_support, opentelemetry_export, audit_logs, rbac, sso_oidc, scim_provisioning, byok_encryption, webhooks_budget_alerts, mcp_gateway [7]
- Regions
- US, EU, India, AWS, GCP, Azure (customer VPC deployments) [8]
- Input types
- chat completions, text completions, embeddings, image generation, audio (speech, transcription, translation), rerank, realtime (WebSocket), fine-tuning, batch, assistants, moderations
- Output types
- streaming (SSE), JSON, OpenAI-compatible response, WebSocket (realtime)
- Webhooks
- ✓ Yes [9]
- Sandbox / test mode
- ✗ No [10]
- SDK languages
- Python, Node.js, Node.js (gateway self-host), Python (OpenAI drop-in), Node.js (OpenAI drop-in) [11]
- MCP server
- ✓ Yes [12]
Trust & compliance
- SOC 2
- SOC 2 Type II [13]
- HIPAA
- ✓ Yes [14]
- GDPR
- ✓ Yes [15]
- ISO 27001
- ✓ Yes [16]
- PCI DSS
- – Unknown [17]
- Published SLA
- ✗ No [18]
- Known restrictions
- Semantic caching available on Production ($49/mo) and Enterprise plans only, Free Developer plan limited to 10k recorded logs/month with no overage, Production plan logs capped at 100k/month; overages at $9 per additional 100k requests (up to 3M), Log retention: 3 days (Developer), 30 days (Production), custom (Enterprise), SOC2/HIPAA/GDPR compliance certificates and custom BAAs available on Enterprise plan only, Private cloud / VPC deployment is Enterprise-only, SCIM provisioning and SSO are Enterprise-only, Prompt templates limited to 3 on free Developer plan (unlimited on Production+), Rate limits on virtual keys available to Enterprise and select Pro customers only [19]
Developer surface
Integration
- API style
- rest
- Base URL
- https://api.portkey.ai/v1
- Version
- v1
- Versioning
- url
- Stability
- ga
- Auth methods
- api_key, jwt
- Error format
- openai-compatible
Adoption & maturity
- Launched
- 2023-01-01
- Notable customers
- Snorkel AI, RVO Health, Haptik, SiteGPT, Snorkel AI, Theories Group, Fontys ICT
Other AI Gateway & LLM Routing APIs
Vercel AI Gateway
"AI Gateway provides a unified API to access hundreds of AI models through a single endpoint, with built-in budgets, usage monitoring, and fallbacks."
Bifrost (Maxim AI)
"The fastest, most resilient, enterprise-grade LLM, MCP, and agent gateway."
Cloudflare AI Gateway
"Connect to any model, dynamically route requests, and manage usage, billing, and logs from one unified gateway."
TrueFoundry AI Gateway
"A unified AI gateway to securely manage and govern AI across 1600+ models with policy control, real-time monitoring, and up to 30% cost reduction."
Helicone
"Open-source LLM observability and monitoring platform for developers" - routes, debugs, and analyzes AI applications with access to 100+ models through one API with built-in observability, automatic fallbacks, and zero markup pricing.
OpenRouter
"The Unified Interface For LLMs" - OpenRouter scouts for the best prices, the lowest latencies, and the highest throughput across dozens of providers, offering a single OpenAI-compatible API with automatic fallback, model routing, and unified billing.
References
- ↑Description: portkey.ai
- ↑Pricing model: portkey.ai · portkey.ai
- ↑Published pricing: portkey.ai
- ↑Free tier: portkey.ai · portkey.ai
- ↑Free tier details: portkey.ai · portkey.ai
- ↑Enterprise plan: portkey.ai
- ↑Supported actions: portkey.ai · portkey.ai · portkey.ai
- ↑Regions: portkey.ai · portkey.ai
- ↑Webhooks: portkey.ai · portkey.ai
- ↑Sandbox: portkey.ai
- ↑SDK languages: portkey.ai
- ↑MCP server: portkey.ai · portkey.ai
- ↑SOC 2: portkey.ai · portkey.ai · portkey.ai
- ↑HIPAA: portkey.ai · portkey.ai
- ↑GDPR: portkey.ai · portkey.ai
- ↑ISO 27001: portkey.ai · portkey.ai
- ↑PCI DSS: portkey.ai
- ↑Published SLA: status.portkey.ai
- ↑Known restrictions: portkey.ai · portkey.ai · portkey.ai
Change history
- 2026-06-21 Capabilities: {} → {"guardrails":true,"self_hosted":true,"observability":true,"spend_controls":tru…
- 2026-06-21 Summary Md: (none) → Portkey is a production infrastructure layer for generative AI teams, providing…
- 2026-06-21 Score Setup Speed: (none) → 85
- 2026-06-21 Score Docs Quality: (none) → 25
- 2026-06-21 Score Procurement Friction: (none) → 100
- 2026-06-21 Score Trust Readiness: (none) → 70
- 2026-06-21 Best For: (none) → Prototypes and side projects - free to start, no sales call, Regulated or enter…
- 2026-06-21 Scoring Methodology: (none) → Scores are computed deterministically from this profile's published, sourced fi…
- 2026-06-21 Score Agent Friendliness: (none) → 65
- 2026-06-21 Score Pricing Transparency: (none) → 100
- 2026-06-21 Llms Txt Present: (none) → Yes
- 2026-06-21 Rendering: (none) → static
- 2026-06-21 Has Structured Data: (none) → Yes
- 2026-06-21 Robots Allows Agents: (none) → Yes
- 2026-06-21 Status Page URL: (none) → https://status.portkey.ai
- 2026-06-21 Docs URL: (none) → https://portkey.ai/docs/introduction/what-is-portkey
- 2026-06-21 Llms Txt URL: (none) → https://portkey.ai/llms.txt
- 2026-06-21 Supported Regions: set to US, EU, India, AWS, GCP, Azure (customer VPC deployments)
- 2026-06-21 Supported Languages: set to (none)
- 2026-06-21 Input Types: set to chat completions, text completions, embeddings, image generation, audio (speech…
- 2026-06-21 Output Types: set to streaming (SSE), JSON, OpenAI-compatible response, WebSocket (realtime)
- 2026-06-21 Webhooks Supported: set to Yes
- 2026-06-21 Sandbox Available: set to No
- 2026-06-21 SDK Languages: set to Python, Node.js, Node.js (gateway self-host), Python (OpenAI drop-in), Node.js …
- 2026-06-21 Enterprise Plan Available: set to Yes
- 2026-06-21 SOC 2: set to type_2
- 2026-06-21 HIPAA: set to Yes
- 2026-06-21 SLA Published: set to No
- 2026-06-21 Data Retention Policy URL: set to https://portkey.ai/privacy-policy
- 2026-06-21 Known Restrictions: set to Semantic caching available on Production ($49/mo) and Enterprise plans only, Fr…
- 2026-06-21 Auth Methods: set to api_key, jwt
- 2026-06-21 Auth Docs URL: set to https://portkey.ai/docs/api-reference/inference-api/authentication
- 2026-06-21 API Style: set to rest
- 2026-06-21 Base URL: set to https://api.portkey.ai/v1
- 2026-06-21 API Version: set to v1
- 2026-06-21 Versioning Scheme: set to url
- 2026-06-21 Stability: set to ga
- 2026-06-21 MCP URL: set to https://mcp.portkey.ai
- 2026-06-21 Quickstart URL: set to https://portkey.ai/docs/guides/getting-started/getting-started-with-ai-gateway
- 2026-06-21 Error Format: set to openai-compatible
- 2026-06-21 Webhook Events URL: set to https://portkey.ai/docs/integrations/guardrails/bring-your-own-guardrails
- 2026-06-21 Requires Verification: set to No
- 2026-06-21 SDK Packages: set to Python, Node.js, Node.js (gateway self-host), Python (OpenAI drop-in), Node.js …
- 2026-06-21 Price Basis: set to month
- 2026-06-21 Free Tier Limit: set to 10,000 recorded logs/month (Dev plan); open-source self-host available (MIT)
- 2026-06-21 Launched At: set to 2023-01-01
- 2026-06-21 Notable Customers: set to Snorkel AI, RVO Health, Haptik, SiteGPT, Snorkel AI, Theories Group, Fontys ICT
- 2026-06-21 Fields Not Found: set to pci_dss, documented_rate_limits, sla_published, minimum_commitment, deprecation…
- 2026-06-21 Source Confidence: set to high
- 2026-06-21 Extractor: set to claude-subagent:sonnet
Suggest an edit / leave a review
Leave a review or comment
curl -X POST https://apio.sh/api/feedback/portkey \
-H 'Content-Type: application/json' \
-d '{"kind":"review","rating":5,"body":"Your experience with this API…"}'Suggest a correction to a field (cite a source)
curl -X POST https://apio.sh/api/suggest/portkey/FIELD \
-H 'Content-Type: application/json' \
-d '{"value":"corrected value","citations":[{"url":"https://source.example/page","excerpt":"supporting quote"}],"note":"what changed and why"}'