Commonstack
Overview
Commonstack is an AI inference marketplace and gateway that provides unified access to multiple LLM providers (OpenAI, Anthropic, Google, DeepSeek, xAI, MiniMax, Zhipu) through a single API. The platform offers quad protocol support (OpenAI Chat Completions, OpenAI Responses API, Anthropic Messages, and Google Gemini endpoints) at the same domain, pay-as-you-go token-based pricing, and targets developers needing multi-provider access without managing separate accounts.
Score: 2/10
Very early-stage competitor. Limited brand recognition, smaller model catalog than OpenRouter but expanding rapidly (now includes GPT-5 series and open-source models). Differentiators include quad SDK compatibility (OpenAI Chat Completions + OpenAI Responses API + Anthropic + Google Gemini), Alipay payment support, and OpenClaw agent integration. The 20% first-deposit bonus suggests aggressive user acquisition but limited organic traction.
Markets
- Developers needing unified multi-provider LLM access
- Chinese-market developers (Alipay support, Zhipu/MiniMax/DeepSeek models)
- AI coding tool users (Claude Code, Cursor, Roo Code integrations)
Products
- Unified API Gateway — single API key for OpenAI, Anthropic, Google, DeepSeek, xAI, MiniMax, Zhipu models
- Quad Protocol Endpoints — OpenAI Chat Completions (
/v1), OpenAI Responses API (/v1/responses), Anthropic Messages, and Google Gemini (/google) at same domain
- API Playground — browser-based model testing
- Team Accounts — shared credits and API key management
- Clawdbot/Clawbox — desktop application with multi-channel integration (Discord, Slack, Telegram) for no-CLI AI access
- UncommonRoute — open-source local smart routing layer with ML-based difficulty scoring, multiple routing modes, and spend controls
- Video Generation API — async video generation via Sora 2, Seedance 2.0, Kling Video O1, and Veo 3.1
Supported Models
| Provider |
Models |
Notes |
| OpenAI |
GPT-5.4, GPT-5.2, GPT-5, GPT-4.1, GPT-OSS-120B |
GPT-5.4 recommended for coding via Responses API; GPT-5.2 is default for OpenClaw; GPT-OSS-120B is open-source 120B model |
| Anthropic |
Claude Sonnet 4.6, Claude Opus 4.6, Claude Opus 4.5 |
1M context |
| Google |
Gemini 3.1 Pro Preview, Gemini 3.1 Flash Lite Preview, Gemini 3 Flash Preview, Gemini 2.5 Flash, Gemini 2.5 Pro, Gemini 3.1 Flash Image Preview, Gemini 2.5 Flash Image, Gemini 3 Pro Image Preview |
Image generation + native Gemini SDK access |
| DeepSeek |
DeepSeek R1, DeepSeek R1 0528 |
— |
| xAI |
Grok 4.1 Fast |
— |
| MiniMax |
MiniMax M2.5 |
205K context |
| Zhipu |
GLM 5 |
203K context |
| Last verified: 2026-04-17 |
|
|
Key Capabilities
| Capability |
Status |
Notes |
| OpenAI SDK compatibility |
Yes |
/v1 endpoint |
| Anthropic SDK compatibility |
Yes |
Same domain, x-api-key auth |
| Google Gemini SDK compatibility |
Yes |
/google endpoint, x-goog-api-key auth; supports gemini-3.1-pro-preview, gemini-3.1-flash-lite-preview, gemini-3-flash-preview, gemini-2.5-flash, gemini-2.5-pro, gemini-3.1-flash-image-preview, gemini-2.5-flash-image, gemini-3-pro-image-preview |
| OpenAI Responses API compatibility |
Yes |
/v1/responses endpoint; supports stateful multi-turn via previous_response_id, reasoning effort (none–xhigh), vision detail levels, structured JSON output |
| Streaming |
Yes |
SSE-based |
| Function calling |
Yes |
— |
| Structured output |
Yes |
output_config on Anthropic endpoint |
| Vision |
Yes |
Image-to-text |
| Image generation |
Yes |
Via Google Gemini image models, OpenAI GPT-4.1, and Nano Banana models |
| Video generation |
Yes |
Sora 2, Seedance 2.0, Kling Video O1, Veo 3.1; supports reference video/audio inputs (Seedance 2.0) |
| PDF processing |
Yes |
— |
| Reasoning/thinking |
Yes |
reasoning param (OpenAI), thinking with budget_tokens (Anthropic) |
| Prompt caching |
Yes |
Cache pricing live for Claude, MiniMax, GLM 5 |
| Web search |
Yes |
Via Anthropic SDK compatibility |
| Fine-tuning |
Yes |
Mentioned on homepage |
| Stream cancellation |
Yes |
— |
| Performance telemetry |
Yes |
Tokens/sec and TTFT in responses |
| Performance tiers |
Yes |
Performative (fast) vs Essential (standard) per model |
| Intelligent routing/failover |
Yes |
UncommonRoute — open-source local routing with ML difficulty scoring, Thompson Sampling model selection; supports Claude Code, Codex, Cursor, OpenClaw |
| Service Access Keys |
Yes |
Server-side credentials for programmatic API key management, chat log queries, usage stats, and balance access |
| Last verified: 2026-04-16 |
|
|
Pricing
| Item |
Price |
Notes |
| Claude Sonnet 4.6 |
$3/$15 per 1M tokens (in/out) |
Cache: $3.75/$0.30 |
| Claude Opus 4.6 |
$5/$25 per 1M tokens (in/out) |
Cache: $6.25/$0.50 |
| MiniMax M2.5 |
$0.30/$1.20 per 1M tokens (in/out) |
Cache read: $0.03 |
| GLM 5 (Zhipu) |
$1/$3.20 per 1M tokens (in/out) |
Cache read: $0.20 |
| Video generation |
$1/1M tokens (USD); ¥1/100K tokens (RMB, ~$1.4/1M) |
Seedance 2.0, Kling Video O1, Veo 3.1 |
| Payment methods |
Credit cards, Alipay |
— |
| New user bonus |
20% on first top-up, up to $500 |
— |
| Last verified: 2026-04-14 |
|
|
URLs to Monitor
| URL |
Label |
Notes |
https://commonstack.ai |
Homepage |
Main landing page |
https://docs.commonstack.ai |
Docs |
Documentation hub |
https://docs.commonstack.ai/overview/quickstart |
Quickstart |
Getting started guide |
https://docs.commonstack.ai/overview/models-and-pricing |
Models & Pricing |
Model catalog and pricing |
https://docs.commonstack.ai/overview/faq |
FAQ |
Frequently asked questions |
https://docs.commonstack.ai/platform/payment-and-billing |
Payment & Billing |
Billing details |
https://docs.commonstack.ai/user-guide/authentication |
Authentication |
Auth documentation |
https://docs.commonstack.ai/user-guide/smart-routing |
Smart Routing (UncommonRoute) |
Open-source local routing layer |
https://docs.commonstack.ai/user-guide/video-generation |
Video Generation |
Video generation API docs |
https://docs.commonstack.ai/user-guide/google-gemini-models-compatibility |
Google Gemini Models Compatibility |
Gemini SDK compatibility docs |
https://docs.commonstack.ai/user-guide/openai-responses-compatible |
OpenAI Responses API Compatibility |
OpenAI Responses API support |
https://docs.commonstack.ai/platform/service-access-key |
Service Access Key |
Server-side key management |
Strategy
- Positioning as an affordable multi-provider gateway with emphasis on Chinese payment methods (Alipay) and Chinese model providers (MiniMax, Zhipu, DeepSeek)
- Aggressive user acquisition via 20% first-deposit bonus (up to $500)
- Building integrations with AI coding tools (Claude Code, Cursor, VS Code, Roo Code), OpenClaw MCP, and Clawdbot/Clawbox desktop app
- Quad SDK compatibility (OpenAI Chat Completions + OpenAI Responses API + Anthropic + Google Gemini) as key differentiator vs single-protocol gateways — now supports OpenAI's newer Responses API alongside Chat Completions
- Performative/Essential two-tier pricing model — differentiated throughput levels per model
- Full API reference documentation now published, signaling platform maturity push
- Prompt caching now live; UncommonRoute smart routing now launched as open-source local proxy
- Video generation API now live with Seedance 2.0, Kling, and Veo 3.1 — expanding beyond text/image into video; Seedance 2.0 supports reference video/audio inputs
- Expanding social presence (Twitter/X: @commonstack_ai)