OpenAI compatible API · Attested · Public status

DeepSeek V4 API Privacy: Attested, No Data to China

Run DeepSeek V4 Pro and V4 Flash on attested, non-Chinese infrastructure. No prompt or output logs. Real-time inference is content-stateless. OpenAI-compatible.

Verify gateway

Onebase URL to migrate

100sof models and routes

0prompt or output logs. Always.

DeepSeek API privacy

DeepSeek V4 on attested infrastructure: your prompts never reach the model vendor.

The worry behind "is the DeepSeek API safe" is specific. api.deepseek.com runs on the vendor's own infrastructure, several US states have restricted it on government devices, and a January 2025 incident exposed plaintext chat logs. The V4 weights themselves are MIT-licensed. The risk lives in where the endpoint runs.

TrustedRouter serves deepseek-v4-pro and deepseek-v4-flash through non-Chinese hosting providers inside a hardware-attested confidential-compute gateway. TrustedRouter never logs prompt or output content. Ordinary synchronous and streaming inference does not retain it. The opt-in Batch API temporarily retains enclave-encrypted artifacts for up to 30 days. Zero-Data-Retention and TEE tiers are available per route. Migration is one line: change base_url, keep your OpenAI SDK.

See DeepSeek pricing Check the attestation

Same SDK, new base_urlPython

from openai import OpenAI

client = OpenAI(
    base_url="https://api.trustedrouter.com/v1",
    api_key=TR_KEY,
)

r = client.chat.completions.create(
    model="deepseek-v4-pro",  # or deepseek-v4-flash
    messages=[{"role": "user", "content": "Summarize this contract."}],
)

Hosting

Served outside China

DeepSeek V4 routes run on non-Chinese hosting providers on attested infrastructure. Prompts do not go to the model vendor. EU-focused routing is available through the EU gateway.

Proof

Attestation you can run

The gateway source is open and the image digest is published. One curl with a nonce returns a JWT signed by the CPU vendor's root key, bound to the live TLS session. Details at /security.

Retention

No prompt or output logs

TrustedRouter never logs prompt or output content. Ordinary synchronous and streaming inference does not retain it. The opt-in Batch API temporarily retains enclave-encrypted artifacts for up to 30 days. Zero-Data-Retention routes add contractual guarantees; TEE routes keep prompts sealed even from the hosting provider. See zero data retention.

api.deepseek.comTrustedRouter

Model accessDeepSeek V4, first-partydeepseek-v4-pro and deepseek-v4-flash, plus 220+ routes across 30+ providers

PricingVendor list prices, hard to beatThin markup over provider list prices; see /pricing

Where prompts goThe vendor's own infrastructureNon-Chinese hosting providers inside an attested gateway

Retention proofPrivacy policy and termsHardware attestation, checkable live, bound to the TLS session

Outage handlingSingle providerAutomatic fallback across providers

Verify, don't trust

Run the attestation check yourself.

Other US hosts also serve DeepSeek outside China with zero-retention policies, and that is a real improvement over first-party routing. They still ask you to take the policy on faith. TrustedRouter's gateway proves it: run curl "https://api.trustedrouter.com/attestation?nonce=$NONCE" and check the returned JWT against the published image digest. Your compliance team gets evidence, plus the DPA, subprocessor list, and SOC 2 readiness documentation to assess it with.

Hosting is the control point for Chinese open-weights generally. In our testing, the same GLM weights answered 60 of 60 sensitive questions through TrustedRouter's enclave versus 30 of 60 via the vendor-hosted API; the filtering is host-applied, not in the weights. One anonymized proof point: a document-processing company scaled from zero to over a billion tokens per week through the attested gateway within three weeks of onboarding. And on DeepSeek R2: unreleased as of July 2026. V4 is what ships today, and it is live on TrustedRouter now. Try it in the playground or compare latency at measured benchmarks.

Live catalog evidence

Current routes, prices, privacy, and measured performance.

Catalog facts come from the routes currently configured in TrustedRouter. Performance uses the same cached metadata snapshot as the public leaderboard. Prompts and outputs are not part of these measurements.

394public models

49providers

1270configured routes

250ZDR routes

8provider E2EE routes

2541recent availability samples

Model	Providers	Context	Input	Output	Privacy	Measured route
DeepSeek: DeepSeek V4 Flash 0731`deepseek/deepseek-v4-flash-0731`	+5 16 routes	1,048,576	$0.0945/1M to $0.147/1M	$0.189/1M to $0.294/1M	ZDR	1193 ms TTFT deepseek · 132 tok/s · 100.00% available · n=187
DeepSeek: DeepSeek V4 Pro`deepseek/deepseek-v4-pro`	+9 24 routes	1,048,576	$0.45675/1M to $2.52/1M	$0.9135/1M to $5.04/1M	ZDR 5 cited scores	3439 ms TTFT gmi · 37 tok/s · 100.00% available · n=75
DeepSeek: DeepSeek V4 Flash 0423`deepseek/deepseek-v4-flash`	+9 22 routes	1,048,576	$0.0945/1M to $0.21/1M	$0.189/1M to $0.42/1M	ZDR 5 cited scores	1787 ms TTFT deepseek · 90 tok/s · 100.00% available · n=49
DeepSeek: DeepSeek V3.2`deepseek/deepseek-v3.2`	+5 15 routes	163,840	$0.273/1M to $1.05/1M	$0.399/1M to $1.575/1M	confidential 3 cited scores	2383 ms TTFT friendli · 100.00% available · n=17
DeepSeek: DeepSeek V3.2 Exp`deepseek/deepseek-v3.2-exp`	5 routes	163,840	$0.2835/1M	$0.4305/1M	varies	4662 ms TTFT siliconflow · 100.00% available · n=6
DeepSeek: DeepSeek V3.1 Terminus`deepseek/deepseek-v3.1-terminus`	7 routes	163,840	$0.2835/1M to $0.315/1M	$0.9975/1M to $1.05/1M	varies	4410 ms TTFT siliconflow · 100.00% available · n=4

DeepSeekPolicy varies 4 models 1193 ms p50 · n=201

DeepInfraPolicy varies 50 models 1704 ms p50 · n=51

Fireworks AIPolicy varies 11 models 2259 ms p50 · n=72

GMI CloudPolicy varies 53 models 3439 ms p50 · n=57

Novita AIPolicy varies 109 models 2073 ms p50 · n=49

SiliconFlowPolicy varies 28 models 2065 ms p50 · n=53

TogetherZDR 18 models 1742 ms p50 · n=35

BasetenPolicy varies 12 models 1767 ms p50 · n=55

Browse every model Review provider policies Open the full leaderboardSnapshot 2026-08-07T03:38:09Z

Questions

Is the DeepSeek API safe to use?

It depends on which endpoint you call. api.deepseek.com is the vendor's own service, and several US states have restricted it on government devices. TrustedRouter serves the same MIT-licensed V4 weights through non-Chinese hosting providers inside a hardware-attested gateway. Prompts do not go to the model vendor. TrustedRouter never logs prompt or output content. Ordinary synchronous and streaming inference does not retain it. The opt-in Batch API temporarily retains enclave-encrypted artifacts for up to 30 days. You can verify the attestation live at any time instead of relying on a policy.

Does using DeepSeek through TrustedRouter send data to China?

No. DeepSeek V4 routes are served by non-Chinese hosting providers on attested infrastructure, so prompts never reach the model vendor. Zero-Data-Retention routes add a contractual guarantee that providers keep nothing, and TEE routes keep the prompt sealed even from the hosting provider. Each route's privacy tier is listed on the models page, and the attestation backing the claim is checkable live.

When will DeepSeek R2 be released?

R2 is unreleased as of July 2026, and no release date is confirmed. The current generation is V4, available as deepseek-v4-pro and deepseek-v4-flash, and both are live on TrustedRouter today. Because the API is OpenAI-compatible, pointing existing code at V4 now and at newer DeepSeek routes later is a one-line model-id change.

How do I migrate off the legacy deepseek-chat alias?

DeepSeek's legacy deepseek-chat alias is scheduled to retire on July 24, 2026. Moving to TrustedRouter takes two edits: set base_url to https://api.trustedrouter.com/v1 and pick a DeepSeek V4 route from the models page. Your OpenAI SDK and the rest of your code stay the same, and automatic provider fallback keeps requests flowing during provider outages.