OFFICIAL APIS ONLY · VERIFIED HOURLY

Every official model.
One key.
More AI, less cost.

GPT, Claude, Gemini, DeepSeek, GLM — and Seedance 2.5 video — routed to the real, official endpoints with public per-model multipliers you can audit. Change one line, keep your SDK.

gpt    = OpenAI(base_url="https://router.flatkey.ai/v1", api_key="sk-fk-…") claude = Anthropic(base_url="https://router.flatkey.ai", api_key="sk-fk-…")

46B tokens/month · 4K+ paying users · 100+ enterprises in production · 0 model swaps — public audit log ↗

LIVE MODEL BOARDALL VERIFIED HOURLY ✓ · REAL 30-DAY TRAFFIC

A
flatkey-auto picks the best official model per request
no routing fee
gpt-5.5openai$5.00$2.00/M in
claude-sonnet-4-6anthropic$3.00$1.60/M in
gpt-5.4openai$2.50$1.00/M in
claude-opus-4-8anthropic$5.00$2.67/M in
deepseek-v4-flashdeepseek$0.14$0.075/M in
glm-5.2z.ai$1.40$0.56/M in
seedance-2.5bytedanceVIDEO · NEW$0.24/s video
FREE START

$2 credit + 3 runs before signup

Together and Segmind killed their free tiers; Novita is down to $0.50. flatkey still lets you run real models before you even sign in.

SMART ROUTING

flatkey-auto, no routing fee

One model name; the gateway picks the best official model per request — task-aware, health-aware, always the lowest multiplier that meets quality.

TRUST · ZERO QUANTIZATION

Official APIs, audited hourly

No fp8 re-serves, no silent swaps. Every model probed against its official fingerprint, on a public log — 160+ models probed, every hour, on the record.

PAY LOCAL

Pix, UPI, Alipay — your currency

Prices in BRL, INR or CNY at checkout, 3DS-secured. Plus Anthropic-shaped API: point Claude Code or Cline at one base_url and go.

Claude at 53% of list.
Everyone else charges full price.

Claude Opus 4.8, $ per 1M input tokens after top-up bonus, measured across gateways on 2026-07-14.

Anthropic official
$5.00
WaveSpeed
$5.00
Replicate
$5.00
Nonelinear
$4.90
flatkey
$2.67

Together, Novita and Segmind don’t carry official Claude at all. flatkey self-serve: 60–90% of list, bonus stacks to an effective 50% — and enterprise interfaces price below self-serve.

RELIABILITY · SLA, IN WRITING

Stability you can
hold us to.

One request, five upstream channel classes. When a channel degrades, the router reroutes automatically — and the SLA is a public document, not a sales promise.

99.5%signed SLA — public terms →
5channel classes, automatic failover
$0balance consumed by flatkey-side errors
Live status →SLA & service credits
UPTIME · SINGLE CHANNEL VS FLATKEY MESH · LIVELIVE
flatkey mesh 99.98% least stable single channel

When one upstream errors, the router recovers on the next-best official channel — you see one flat line.

Why choose
flatkey?

As low as 50% of list

Models at 60–90% of list; the top-up bonus stacks to an effective 50%. One prepaid balance across every model.

official
$5.00
flatkey
$2.67

claude-opus-4-8 · $/1M input tokens, after bonus

Official endpoints, verified hourly

Every request hits the real lab API — no fp8 re-serves, no silent swaps. 160+ models probed against official fingerprints, on a public log.

GOpenAI
CAnthropic
GGoogle
DDeepSeek
QAlibaba
ZZ.ai

One SDK, every modality

Text, video, audio and image behind the same OpenAI-compatible gateway — and an Anthropic-shaped API for Claude Code.

client = OpenAI(
  base_url="https://router.flatkey.ai/v1")
client.chat.completions.create(
  model="seedance-2.5",
  messages=[…])  # video, same SDK

Enterprise-grade governance

Tree-structured sub-keys with budgets and model allowlists, a per-request ledger API, invoices in 48h — on audited infrastructure.

Sub-key capsModel allowlistsLedger APIInvoices 48hSOC 2 Type IIISO 27001GDPRZero retention
TRUST · HOURLY MODEL-AUTHENTICITY PROBES

0 model swaps. Ever.
And we prove it.

Every hour, every model is probed against its official API fingerprint — logits, tokenizer, refusal shape. The log is public. Cheap gateways ask for trust; flatkey shows receipts.

14:00claude-opus-4-8 fingerprint ✓ · official anthropic · 388msPASS
14:00gpt-5.5 fingerprint ✓ · official openai · 412msPASS
14:00glm-5.2 fingerprint ✓ · official zhipu · 241msPASS
14:00seedance-2.5 fingerprint ✓ · official bytedance · videoPASS
13:00160+ models probed hourly · 30-day log publicPASS
BACKGROUND VIDEO · hourly-probe radar · 12s loop
req_8f3a2e91 · in 1,204 tok · out 388 tok · cached 61% · $0.0041
req_c11977ab · in 640 tok · out 1,022 tok · cached 38% · $0.0009
req_40deb0c2 · in 8,441 tok · out 96 tok · cached 95% · $0.0002
invoice #2026-0714 · VAT · issued 48h
req_77ab40de · 3DS ✓ · stripe checkout · BRL R$55,00
audit_export.csv · 128,441 rows · sha256 ✓
subkey cline-agent · cap $50/mo · model allowlist ✓
subkey prod · cap $500/mo · alert 80% ✓
SLA 99.9% · failover 3-region · breaker 99.95%
gdpr export · user 1835 · 2.1MB · done
req_b0c240de · in 96 tok · out 2,048 tok · $0.0031
COMPLIANCE · ENTERPRISE-GRADE AUDITABILITY

Every token,
on the record.

Per-request ledger with input, output and cached tokens — exportable, API-accessible. Invoices in 48h (VAT fapiao supported), 3DS payments, sub-key caps and model allowlists for teams.

99.5% SLAmulti-provider failover · signed
3DS · Stripefraud-screened payments
Invoices 48henterprise invoicing API
Ledger APIevery request, every token
Sub-key governancecaps · allowlists · alerts
BACKGROUND VIDEO · live ledger stream · 15s loop
gpt-5.5✓ verified · openai · 1M$0.75/M · 60% of list
claude-opus-4-8✓ verified · anthropic · 500K$2.50/M · 50% of list
gemini-3.2-pro✓ verified · google · 2M$1.50/M
deepseek-v4✓ verified · deepseek · 256K$0.17/M
qwen3-maxalibaba$0.72/M
glm-5.2✓ verified · zhipu · 200K$0.36/M
gpt-5.5✓ verified · openai · 1M$0.75/M · 60% of list
claude-opus-4-8✓ verified · anthropic · 500K$2.50/M · 50% of list
gemini-3.2-pro✓ verified · google · 2M$1.50/M
deepseek-v4✓ verified · deepseek · 256K$0.17/M
qwen3-maxalibaba$0.72/M
glm-5.2✓ verified · zhipu · 200K$0.36/M
seedance-2.5✓ verified · bytedance · video$0.24/s · t2v + i2v
kimi-k3moonshot$0.55/M
claude-sonnet-5✓ verified · anthropic · 1M$1.80/M
gpt-5.4-mini✓ verified · openai · 400K$0.30/M
gemini-3-flashgoogle$0.05/M
doubao-seed-2bytedance$0.16/M
ernie-5.1baidu$0.55/M
seedance-2.5✓ verified · bytedance · video$0.24/s · t2v + i2v
kimi-k3moonshot$0.55/M
claude-sonnet-5✓ verified · anthropic · 1M$1.80/M
gpt-5.4-mini✓ verified · openai · 400K$0.30/M
gemini-3-flashgoogle$0.05/M
doubao-seed-2bytedance$0.16/M
ernie-5.1baidu$0.55/M
flatkey-autoroutes itselfno routing fee
minimax-m3minimax$0.18/M
grok-4.2xai$1.80/M
step-3.7-flashstepfun$0.16/M
hunyuan-2.0tencent$0.40/M
mimo-v2.5xiaomi$0.01/M
flatkey-autoroutes itselfno routing fee
minimax-m3minimax$0.18/M
grok-4.2xai$1.80/M
step-3.7-flashstepfun$0.16/M
hunyuan-2.0tencent$0.40/M
mimo-v2.5xiaomi$0.01/M
MODELS · 160+ OFFICIAL MODELS, ONE KEY

Every frontier lab.
One key.

OpenAI, Anthropic, Google, DeepSeek, Alibaba, Zhipu, Moonshot, ByteDance… new models live within 24h of official release — always the real endpoints, never quantized re-serves.

BACKGROUND VIDEO · model wall marquee · 30s loop
VIDEO MODELS · SAME KEY, SAME BALANCE

Video generation, official endpoints.

Text-to-video and image-to-video from every major lab — metered per second on the same prepaid balance as your text models.

sora-2openai · t2v$0.50/s
kling-2.5-prokuaishou · i2v$0.21/s
wan-2.7alibaba · t2v$0.12/s
hailuo-02minimax · i2v$0.14/s
pixverse-v6pixverse · t2v$0.10/s
seedance-2.0-litebytedance · t2v$0.08/s
ltx-2.3lightricks · i2v$0.06/s
AUDIO MODELS · VOICE IN, VOICE OUT

Speech, on the record too.

Realtime voice, TTS and transcription from the official labs — same key, same ledger, priced per minute or per million characters.

gpt-realtimeopenai · speech ⇄ speech$0.06/min
whisper-v4openai · transcription$0.003/min
elevenlabs-v3-turboelevenlabs · tts$0.08/1K chars
minimax-speech-2.6minimax · tts$0.02/1K chars
cosyvoice-3alibaba · tts$0.01/1K chars
seed-tts-2bytedance · tts$0.015/1K chars
gemini-3-audiogoogle · native audio$0.04/min

“We route support and product-copy workloads across GPT and Claude through one flatkey balance. Same SDK we already had — roughly half the bill.”

AAI Platform TeamAnker Innovations

“The hourly authenticity probes ended the ‘is this the real model?’ debate. Procurement gets one invoice; engineers get official endpoints.”

LSmart-Home AI LeadLockin

Built on flatkey,
loved by builders.

“Compliance was the blocker until flatkey: zero retention of request content, SOC 2 controls, and a per-request ledger. Our health-data reviews finally pass — and the bill dropped 47% in the first month.”

iHData & AI LeadiHealth Labs
HOW IT FITS TOGETHER

From homepage to production calls.

1

Get one key

Create a flatkey account, open the console, and generate an API key for your app.

2

Change the base URL

Point your existing SDK — OpenAI or Anthropic protocol, GPT, Claude, Gemini alike — at router.flatkey.ai. Nothing else changes.

3

Monitor and optimize

Review usage, cost, routing and errors from the same dashboard — per model, per token, per request.

SUPPORT · HUMANS ON CALL

Questions? Talk to us.

Integration, model choice, or billing — online support and usage consultation for everything on flatkey.

More AI, less cost —
on every official model

Whether you ship an agent today or route billions of tokens a month — one key, a signed SLA, and prices that drop as you grow.