Vercel AI Gateway

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 model

Usage-based [2]vercel.com/docs/ai-gateway/pricing“AI Gateway uses a pay-as-you-go model with no markups. Purchase AI Gateway Credits and Vercel automatically deducts charges from your balance.”vercel.com/docs/ai-gateway/pricing“For the paid tier, AI Gateway provides tokens with zero markup, including when you bring your own key.”

Published pricing

✓ Yes [3]vercel.com/docs/ai-gateway/pricing“Custom Reporting: Write $0.075 / 1,000 tag/user ID/quota entity ID writes; Query $5 / 1,000 queries to the reporting endpoint... Team-wide provider allowlist $0.10 per 1,000 successful requests... Team-wide zero data retention $0.10 per 1,000 requests”

Free tier

✓ Yes [4]vercel.com/ai-gateway“When you sign up for a Vercel account, you get $5 of credits every 30 days to try out any model from our model list.”vercel.com/docs/ai-gateway/pricing“Every Vercel team account gets access to both a free tier and a paid tier for AI Gateway Credits. The free tier includes a subset of models, not the full catalog. Free tier requests are also rate limited per model, with lower limits than the paid tier.”

Free tier details

$5 of AI Gateway Credits every 30 days for teams that have not purchased credits. Free tier covers a subset of models (not the full catalog) and is rate-limited per model. Monthly free credit stops applying once the team makes a first purchase of AI Gateway Credits. [5]vercel.com/docs/ai-gateway/pricing“Your free credits start when you make your first AI Gateway request. To run larger workloads, you can purchase AI Gateway Credits at any time with no obligation to renew. Once you purchase credits, your account transitions to the paid tier and the monthly free credit no longer applies.”vercel.com/ai-gateway“When you sign up for a Vercel account, you get $5 of credits every 30 days to try out any model from our model list.”

Self-serve signup

✓ Yes [6]vercel.com/docs/ai-gateway/authentication-and-byok/authentication“Go to the AI Gateway API Keys page in your Vercel dashboard. Click Create key and follow the steps to generate a new API key.”

Requires sales call

✗ No

Enterprise plan

✓ Yes [7]vercel.com/docs/ai-gateway/pricing“Team-wide provider allowlist $0.10 per 1,000 successful requests Pro and Enterprise... Zero Data Retention: Per-request zero data retention No additional cost Pro and Enterprise”

• Fallback / routing
• Spend controls
• Observability

Supported actions

unified_chat_completions, openai_compatible_api, anthropic_messages_api_compatible, openai_responses_api_compatible, model_routing, automatic_fallback, load_balancing, prompt_caching, spend_limits, budgets, rate_limiting, observability_logging, tracing, byo_provider_keys, virtual_keys, provider_allowlist, zero_data_retention, disallow_prompt_training, routing_rules, model_rewrite_rules, model_deny_rules, custom_reporting, spend_reporting_by_user_model_tag, embeddings, reranking, image_generation, video_generation, speech_tts_stt, realtime_speech_to_speech, web_search_augmentation, dynamic_model_discovery, oidc_authentication, provider_timeout_config, service_tier_selection [8]vercel.com/docs/ai-gateway/models-and-providers/provider-options“You can use order, only, and sort in providerOptions.gateway to control which providers handle your requests, in what order, and how they are ranked.”vercel.com/docs/ai-gateway/models-and-providers“const rerankingModels = models.filter((m) => m.modelType === 'reranking');”vercel.com/docs/ai-gateway/models-and-providers/routing-rules“Routing rules let you control how AI Gateway handles your team's requests at the model level. You define team-wide rules that either rewrite a request from one model to another, or deny a model so requests for it are blocked.”

Input types

chat completions, text completions, embeddings, reranking, image generation, video generation, audio/speech, realtime speech-to-speech [9]vercel.com/docs/ai-gateway/models-and-providers“const rerankingModels = models.filter((m) => m.modelType === 'reranking');”vercel.com/docs/ai-gateway/getting-started“Text Generation: Generate and stream text... Image Generation: Create images from text prompts... Video Generation: Create videos from text prompts... Realtime: Build low-latency, speech-to-speech voice agents... Speech: Generate spoken audio from text and transcribe audio back to text”

Output types

streaming (SSE), JSON, OpenAI-compatible response, Anthropic-compatible response, OpenAI Responses API-compatible response

Webhooks

✓ Yes [10]vercel.com/docs/spend-management“You can configure a webhook URL to trigger events such as serving a static version of your site, pausing a project, or sending a Slack notification. Vercel will send a HTTPS POST request to the URL that you provide when the following events happen: When a spend amount reaches 100%; At the end of your billing cycle”

Sandbox / test mode

✗ No

SDK languages

TypeScript/JavaScript, TypeScript/JavaScript (OpenAI-compatible drop-in), Python (OpenAI-compatible drop-in), TypeScript (Anthropic-compatible drop-in) [11]vercel.com/docs/ai-gateway/sdks-and-apis/ai-sdk“The AI SDK is the recommended way to build AI-powered TypeScript applications with AI Gateway. Pass a model string like 'anthropic/claude-sonnet-4.6' directly to AI SDK functions and requests route through AI Gateway automatically.”vercel.com/docs/ai-gateway/getting-started“AI Gateway works with AI SDK v5 and v6, OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, or your preferred framework.”

MCP server

✓ Yes [12]vercel.com/docs/agent-resources/vercel-mcp“Vercel MCP is Vercel's official MCP server. It's a remote MCP with OAuth that gives AI tools secure access to your Vercel projects available at: https://mcp.vercel.com”

SOC 2

SOC 2 Type II [13]vercel.com/security“SOC 2 Type 2 compliant. Vercel has obtained a SOC 2 Type 2 attestation”

HIPAA

✓ Yes [14]vercel.com/security“Vercel supports HIPAA compliance for enterprise customers. Their HIPAA report is available upon request at security.vercel.com”

GDPR

✓ Yes [15]vercel.com/security“Vercel confirms GDPR compliance. The platform provides a Data Processing Addendum describing Technical and Organizational Security Measures.”

ISO 27001

✓ Yes [16]vercel.com/security“Vercel holds ISO 27001:2013 certified status.”

PCI DSS

✓ Yes [17]vercel.com/security“Vercel maintains Self-Assessment Questionnaire (SAQ)-D Attestation of Compliance (AOC) for Service Providers and a Self-Assessment Questionnaire (SAQ)-A Attestation of Compliance (AOC) for Merchants based on PCI DSS v4.0.”

Published SLA

✓ Yes [18]vercel.com/legal/sla“Vercel Inc. will use commercially reasonable efforts to make the platform as a service...available 99.99% of the time, excluding any Excused Downtime events.”

Rate limits

Free tier requests are rate-limited per model with lower limits than the paid tier; a 429 is returned when exceeded. Purchasing AI Gateway Credits moves the team to the paid tier with higher rate limits. Vercel states it places no rate limits of its own on paid-tier queries - upstream provider limits may apply. Specific numeric RPM/TPM figures are not publicly documented. [19]vercel.com/ai-gateway“Vercel doesn't place any rate limits on your queries — while the upstream providers may have limits, we actively work with them to maximize your limits and throughput.”vercel.com/docs/ai-gateway/pricing“Free tier requests are also rate limited per model, with lower limits than the paid tier. If you exceed a limit, AI Gateway returns a 429 error and you can retry after a short wait. Purchasing AI Gateway Credits moves your team to the paid tier, which raises your rate limits.”

Known restrictions

BYOK is only available on the paid tier (requires purchased AI Gateway Credits), Per-request zero data retention (ZDR) and provider allowlist are Pro and Enterprise only, Team-wide provider allowlist incurs $0.10 per 1,000 successful requests surcharge (Pro and Enterprise), Team-wide Zero Data Retention incurs $0.10 per 1,000 requests surcharge (Pro and Enterprise), Custom Reporting API endpoint (/v1/report) is not available on Hobby or Pro-trial plans, Custom Reporting incurs $0.075 per 1,000 tag/user ID/quota entity ID writes and $5 per 1,000 queries, Routing rules feature is in beta and may change before GA, Free tier limited to subset of models, not full catalog, Free monthly credit ($5) ceases once the first credit purchase is made [20]vercel.com/docs/ai-gateway/sdks-and-apis/rest-api“Hobby and Pro-trial plans cannot use this endpoint.”vercel.com/docs/ai-gateway/authentication-and-byok/byok“BYOK is available on the paid tier. When a request with your credentials fails, AI Gateway retries it with system credentials for reliability, and that fallback usage is charged against your credits balance. To use your own provider keys, you'll need purchased AI Gateway Credits.”vercel.com/docs/ai-gateway/models-and-providers/routing-rules“AI Gateway routing rules are in beta and may change before general availability. Avoid relying on them in production.”

• Docs
• API reference
• llms.txt
• Changelog
• Status page

Docs rendering: static · llms.txt present

API style

rest

Base URL

https://ai-gateway.vercel.sh/v1

Version

v1

Versioning

url

Stability

ga

Auth methods

api_key, jwt

Error format

openai-compatible

SDKs

• TypeScript/JavaScript ai · repo
• TypeScript/JavaScript @ai-sdk/gateway · repo
• TypeScript/JavaScript (OpenAI-compatible drop-in) openai · repo
• Python (OpenAI-compatible drop-in) openai · repo
• TypeScript (Anthropic-compatible drop-in) @anthropic-ai/sdk · repo

• Auth docs
• Quickstart
• MCP server
• Deprecation policy

Launched

2025-05-20

GA

2025-08-21

Notable customers

v0.app, Browserbase, Perplexity

• Portkey

  "Production Stack for Gen AI Builders"

  Hybrid · free tier · public pricing · self-serve

• Bifrost (Maxim AI)

  "The fastest, most resilient, enterprise-grade LLM, MCP, and agent gateway."

  Sales-led · free tier · public pricing · self-serve

• Cloudflare AI Gateway

  "Connect to any model, dynamically route requests, and manage usage, billing, and logs from one unified gateway."

  Hybrid · free tier · public pricing · self-serve

• 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."

  Hybrid · free tier · public pricing · self-serve

• 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.

  Hybrid · free tier · public pricing · self-serve

• 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.

  Usage · free tier · public pricing · self-serve

Vercel AI Gateway alternatives · Vercel AI Gateway vs Portkey · All AI Gateway & LLM Routing APIs APIs

Each field above carries a numbered source - hover for a preview, click to jump here.

1. ↑Description: vercel.com
2. ↑Pricing model: vercel.com · vercel.com
3. ↑Published pricing: vercel.com
4. ↑Free tier: vercel.com · vercel.com
5. ↑Free tier details: vercel.com · vercel.com
6. ↑Self-serve signup: vercel.com
7. ↑Enterprise plan: vercel.com
8. ↑Supported actions: vercel.com · vercel.com · vercel.com
9. ↑Input types: vercel.com · vercel.com
10. ↑Webhooks: vercel.com
11. ↑SDK languages: vercel.com · vercel.com
12. ↑MCP server: vercel.com
13. ↑SOC 2: vercel.com
14. ↑HIPAA: vercel.com
15. ↑GDPR: vercel.com
16. ↑ISO 27001: vercel.com
17. ↑PCI DSS: vercel.com
18. ↑Published SLA: vercel.com
19. ↑Rate limits: vercel.com · vercel.com
20. ↑Known restrictions: vercel.com · vercel.com · vercel.com

Every field change, who made it, and when - from our audited data pipeline and editors.

1. 2026-06-21 Capabilities: {} → {"observability":true,"spend_controls":true,"fallback_routing":true}
2. 2026-06-21 Summary Md: (none) → Vercel AI Gateway is a unified REST API that routes requests to hundreds of AI …
3. 2026-06-21 Score Agent Friendliness: (none) → 55
4. 2026-06-21 Score Pricing Transparency: (none) → 75
5. 2026-06-21 Score Setup Speed: (none) → 85
6. 2026-06-21 Score Docs Quality: (none) → 55
7. 2026-06-21 Score Procurement Friction: (none) → 90
8. 2026-06-21 Score Trust Readiness: (none) → 100
9. 2026-06-21 Best For: (none) → Prototypes and side projects - free to start, no sales call, Regulated or enter…
10. 2026-06-21 Scoring Methodology: (none) → Scores are computed deterministically from this profile's published, sourced fi…
11. 2026-06-21 Llms Txt URL: (none) → https://vercel.com/llms.txt
12. 2026-06-21 Rendering: (none) → static
13. 2026-06-21 Has Structured Data: (none) → No
14. 2026-06-21 Robots Allows Agents: (none) → Yes
15. 2026-06-21 API Reference URL: (none) → https://vercel.com/login?next=%2Freference
16. 2026-06-21 Status Page URL: (none) → https://status.vercel.com
17. 2026-06-21 Changelog URL: (none) → https://vercel.com/changelog
18. 2026-06-21 Docs URL: (none) → https://vercel.com/docs
19. 2026-06-21 Llms Txt Present: (none) → Yes
20. 2026-06-21 Free Tier Details: set to $5 of AI Gateway Credits every 30 days for teams that have not purchased credit…
21. 2026-06-21 Self Serve Signup: set to Yes
22. 2026-06-21 Requires Sales Call: set to No
23. 2026-06-21 Enterprise Plan Available: set to Yes
24. 2026-06-21 SOC 2: set to type_2
25. 2026-06-21 HIPAA: set to Yes
26. 2026-06-21 GDPR: set to Yes
27. 2026-06-21 ISO 27001: set to Yes
28. 2026-06-21 PCI DSS: set to Yes
29. 2026-06-21 SLA Published: set to Yes
30. 2026-06-21 SLA URL: set to https://vercel.com/legal/sla
31. 2026-06-21 Data Retention Policy URL: set to https://vercel.com/docs/ai-gateway/security-and-compliance/zdr
32. 2026-06-21 Documented Rate Limits: set to Free tier requests are rate-limited per model with lower limits than the paid t…
33. 2026-06-21 Known Restrictions: set to BYOK is only available on the paid tier (requires purchased AI Gateway Credits)…
34. 2026-06-21 Auth Methods: set to api_key, jwt
35. 2026-06-21 Auth Docs URL: set to https://vercel.com/docs/ai-gateway/authentication-and-byok/authentication
36. 2026-06-21 API Style: set to rest
37. 2026-06-21 Base URL: set to https://ai-gateway.vercel.sh/v1
38. 2026-06-21 API Version: set to v1
39. 2026-06-21 Versioning Scheme: set to url
40. 2026-06-21 Stability: set to ga
41. 2026-06-21 Deprecation Policy URL: set to https://vercel.com/docs/release-phases
42. 2026-06-21 MCP URL: set to https://mcp.vercel.com
43. 2026-06-21 Quickstart URL: set to https://vercel.com/docs/ai-gateway/getting-started/text
44. 2026-06-21 Error Format: set to openai-compatible
45. 2026-06-21 Requires Verification: set to No
46. 2026-06-21 Price Basis: set to token
47. 2026-06-21 Slug: set to vercel-ai-gateway
48. 2026-06-21 Launched At: set to 2025-05-20
49. 2026-06-21 GA Date: set to 2025-08-21
50. 2026-06-21 Notable Customers: set to v0.app, Browserbase, Perplexity

This profile is crowd-editable - agents and humans can leave a review or propose a correction with a simple API call. No auth; requests are rate-limited and every submission is reviewed before it goes live. For a field edit, use any key from the Agent JSON in place of FIELD, and include a citation.

curl -X POST https://apio.sh/api/feedback/vercel-ai-gateway \
  -H 'Content-Type: application/json' \
  -d '{"kind":"review","rating":5,"body":"Your experience with this API…"}'

curl -X POST https://apio.sh/api/suggest/vercel-ai-gateway/FIELD \
  -H 'Content-Type: application/json' \
  -d '{"value":"corrected value","citations":[{"url":"https://source.example/page","excerpt":"supporting quote"}],"note":"what changed and why"}'

All the ways to contribute →