De
DeepSeek Chat
deepseek-chatA strong default for coding help, chat workflows, and everyday product features.
Coding and general chat$0.14 in / $0.28 out
OpenAI-compatible API gateway
One API key.
Many models.
Zero rewrites.
Point your OpenAI SDK at one endpoint, switch models without changing code, and pay in USD.
from openai import OpenAI
client = OpenAI(
api_key="sk-...",
base_url="https://api.glideflowai.com/v1",
)
r = client.chat.completions.create(
model="deepseek-chat",
messages=[{"role": "user", "content": "Hello"}],
)https://api.glideflowai.com/v1
OpenAI
Compatible API
USD
Billing, no extra accounts
1 line
To switch models
Stream
Server-sent events
Models worth shipping on
Strong open and frontier-class models for coding, reasoning, and high-volume work. One key reaches all of them.
De
DeepSeek Reasoner
deepseek-reasonerStep-by-step reasoning for debugging, planning, and harder multi-step tasks.
Reasoning and debugging$0.55 in / $2.19 out
Qw
Qwen Max
qwen-maxA capable multilingual model for larger prompts and polished product experiences.
Multilingual apps$0.40 in / $1.20 out
GL
GLM-4
glm-4A practical low-cost option for prototypes, agents, and high-volume background jobs.
Low-cost volume$0.20 in / $0.60 out
Built for switching
Change the model, not your code
Every model shares the same request shape and the same key. Test a cheaper model, swap to a stronger one, and keep the rest of your app exactly as it is.
- One endpoint, many models
Keep the OpenAI request shape you already use. Change the model id, not your integration.
- Routing with fallback
Each model can fail over to a backup provider, so a single upstream hiccup does not take you down.
- Usage you can read
Per-request logs show the model, tokens, latency, and cost. No guessing where the bill came from.
# one client, many models
for model in [
"deepseek-chat",
"qwen-max",
"glm-4",
]:
client.chat.completions.create(
model=model,
messages=msgs,
)Price and performance, side by side
The same OpenAI request shape, at open-model economics. Compare per-million-token pricing against frontier list prices.
| Provider | Model | Input | Output | Note |
|---|---|---|---|---|
| Glideflow | DeepSeek Chat | $0.14 / 1M | $0.28 / 1M | Launch placeholder |
| Glideflow | DeepSeek Reasoner | $0.55 / 1M | $2.19 / 1M | Launch placeholder |
| OpenAI | GPT-5.5 | $1.75 / 1M | $14 / 1M | Public list price |
| Anthropic | Claude Sonnet 4.6 | $3 / 1M | $15 / 1M | Public list price |
| Gemini 2.5 Pro | $1.25 / 1M | $10 / 1M | Public list price, lower context tier |
Per-model pricing
| Model | Best for | Input | Output |
|---|---|---|---|
DeepSeek Chatdeepseek-chat | Coding and general chat | $0.14 / 1M tokens | $0.28 / 1M tokens |
DeepSeek Reasonerdeepseek-reasoner | Reasoning and debugging | $0.55 / 1M tokens | $2.19 / 1M tokens |
Qwen Maxqwen-max | Multilingual apps | $0.40 / 1M tokens | $1.20 / 1M tokens |
GLM-4glm-4 | Low-cost volume | $0.20 / 1M tokens | $0.60 / 1M tokens |
From zero to first call
01
Create your account
Sign up with email in under a minute. No company details, no sales call.
02
Add credit
Top up in USD with a redeem code. Your balance is shared across every model.
03
Call the API
Point your OpenAI SDK at one base URL and ship. Switching models is one line.
Drop it into your stack
Three ways in, all OpenAI-compatible. Copy, paste your key, ship.
curl https://api.glideflowai.com/v1/chat/completions \
-H "Authorization: Bearer sk-xxx" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-chat",
"messages": [{"role": "user", "content": "Hello"}]
}'from openai import OpenAI
client = OpenAI(
api_key="sk-xxx",
base_url="https://api.glideflowai.com/v1",
)
r = client.chat.completions.create(
model="deepseek-chat",
messages=[{"role": "user", "content": "Hello"}],
)
print(r.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
apiKey: "sk-xxx",
baseURL: "https://api.glideflowai.com/v1",
});
const r = await client.chat.completions.create({
model: "deepseek-chat",
messages: [{ role: "user", content: "Hello" }],
});
console.log(r.choices[0].message.content);Clear about where requests go
Glideflow is for developers who value cost and compatibility, and who choose what data is right for third-party routing.
Transparent routing
Requests reach the Glideflow gateway, get authenticated and metered, then forward to the model provider you choose.
Built for indie workloads
Prototypes, coding tools, side projects, and early products where practical cost beats enterprise procurement.
Not for regulated data
Do not send medical, financial, or government data unless your own review approves the full provider chain.
Straight answers
Is it really OpenAI-compatible?+
Yes. Use the official OpenAI SDKs or plain HTTP, set the base URL to ours, and call chat completions and models the same way.
How do payments work?+
Create an account, then top up your balance in USD with a redeem code. No Chinese phone number, local account, or identity flow is required.
How is this different from running each provider myself?+
One account, one key, and one balance cover every model. You skip registering, funding, and integrating each provider separately.
What happens if a model is unavailable?+
A model can route to a backup provider, and your usage logs record which provider served each request.
Which SDKs are supported?+
Any OpenAI-compatible client works. The docs include curl, the Python openai SDK, and the Node openai SDK.
Ship with one model API today
Create an account, add credit, and make your first call in minutes.