Added standalone API presets: 31 image request templates and 8 video presets. Current-key pricing matches 26 image and 8 video presets; 5 image templates block new paid calls while pricing is unavailable. Native provider/UI/import/render acceptance remains as documented in the original guide below.
| Image model / tested preset | Current preset quote / image | Historical example amount / image | Fee evidence |
|---|---|---|---|
| gpt-image-2(VIP) low |
$0.0019/image | $0.0019/image | Billing log, phase unknown |
| gpt-image-2.5-flare-vip low |
Current-key quote unavailable | $0.005272/image Token-metered; current $/M unavailable. This sample is not a fixed image tariff. |
Task receipt settled |
| nano_banana 1K |
$0.012/image | $0.012/image | Billing log, phase unknown |
| gemini-nano-banana-2.1 1K |
$0.015/image | $0.015/image | Billing log, phase unknown |
| gpt-image-2 1K |
$0.015/image | $0.015/image | Billing log, phase unknown |
| gemini-nano-banana-2.1-vip 1K |
$0.02/image | $0.02/image | Billing log, phase unknown |
| qwen-image-3.0 1K |
Current-key quote unavailable | $0.021/image | Billing log, phase unknown |
| gpt-image-2.5-flare high |
$0.0225/image | $0.0225/image | Billing log, phase unknown |
| gpt-image-2.5-sunburst high |
$0.0225/image | $0.0225/image | Billing log, phase unknown |
| flux-2-pro 1K |
$0.025/image | $0.025/image | Billing log, phase unknown |
| doubao-seedream-5-0-pro 1K |
$0.042857/image | $0.042856/image | Billing log, phase unknown |
| Video model | Preset / input | Settled clip charge | File duration | Case USD/s, ascending |
|---|---|---|---|---|
| MiniMax-H3-Max-Turbo | 5s / 480P / text-to-video | $0.125/clip | 5.18s | $0.024131/s Pricing tariff: $0.025/s |
| seedance-2-mini | 4s / 480p / first-frame | $0.13338/clip | 4.04s | $0.033015/s Pricing tariff: $3.285714/M total tokens |
| seedance-2-fast | 4s / 480p / text-to-video | $0.214568/clip | 4.04s | $0.053111/s Pricing tariff: $5.285714/M total tokens |
| kling-v3 | 5s / 720p / std / text-to-video | $0.3/clip | 5.04s | $0.059524/s Pricing tariff: $0.06/s |
| MiniMax-H3 | 5s / 768p / text-to-video | $0.32/clip | 5.17s | $0.061896/s Pricing tariff: $0.064/s |
| seedance-2 | 4s / 480p / first-frame | $0.26676/clip | 4.04s | $0.06603/s Pricing tariff: $6.571429/M total tokens |
| kling-3.0-turbo | 4s / 720p / text-to-video | $0.32/clip | 4.04s | $0.079208/s Pricing tariff: $0.08/s |
| seedance-2-5 | 4s / 480p / text-to-video | $0.3883/clip | 4.04s | $0.096114/s Pricing tariff: $10/M total tokens |
Pricing checked 2026-10-20. Image examples are dated 2026-10-17; video examples 2026-10-20. Tables sort by example amount or case USD/s. Presets and input modes differ; this is not a matched-settings quality ranking. Unknown-phase image logs are not confirmed final net settlements. Current quotes and historical amounts are separate. Video charges are billing=settled; case USD/s divides charge by actual file duration, never processing time. Token tariffs appear in small text and do not imply a fixed per-second price.
gpt-image-2(VIP) starts at $0.0019/image; this low/1K preset is $0.0019/image. The standalone sampler has 31 historically successful image request templates and 8 video presets. Current-key pricing matches 26 image and 8 video presets; 5 image templates block new paid submission until an exact price is available.
Local sampler · Full price table · Original media and fee evidence
Original project examples, historical charges and integration instructions follow. Historical amounts are not rewritten as today’s quotes.
One Python script, three API workflows. Call a chat model, create an image, or submit a video generation task through ToAPIs. No third-party Python packages are required.
Built for developers connecting models to applications and recurring workflows. Start with one verified request, inspect actual usage, then plan your production integration. High-volume integration enquiries should include your models, estimated monthly usage, peak concurrency, and launch timeline.
Start with
--dry-run: it shows the endpoint and request body without an API key or billable call. Live calls require a key and may incur charges. Model availability and prices can change; check the current pricing page.
Tested on Oct 12: MeiGen local MCP with a community patch and TongFlow plugin image-gen CLI. Each guide includes a real PNG, pinned version, reproduction steps and dated pricing.
| Workflow | Model used in this example | API behavior |
|---|---|---|
| Text | gpt-6.1-sol |
Synchronous chat completion |
| Image | gpt-image-2-vip (1K, low) |
Submit task, then poll for result URL |
| Video | seedance-2 (4 s, 720p, no audio) |
Submit task, then poll for result URL |
The script follows the official quickstart. ToAPIs also lists Suno music generation; this quickstart has no Suno example because its request and task response still need a verified API contract. It does not cover TTS.
Install Python 3.10 or newer, clone this repository, then run:
python examples/quickstart.py chat --dry-run
python examples/quickstart.py image --prompt "A paper crane on an orange background" --dry-run
python examples/quickstart.py video --prompt "A paper crane unfolds on a desk, cinematic light" --dry-runOn Windows, py may be used in place of python.
Create a ToAPIs API key in your account. Store it as an environment variable. Do not paste it into a source file or commit it to GitHub.
macOS / Linux:
export TOAPIS_API_KEY="YOUR_API_KEY"
python examples/quickstart.py chat --prompt "Explain an API gateway in one sentence."PowerShell:
$env:TOAPIS_API_KEY = "YOUR_API_KEY"
python .\examples\quickstart.py chat --prompt "Explain an API gateway in one sentence."For users served by the China endpoint, set TOAPIS_BASE_URL=https://toapis.cn (or pass --base-url https://toapis.cn). The script appends /v1 automatically.
python examples/quickstart.py image --prompt "A clean orange-and-black geometric poster"
python examples/quickstart.py video --prompt "A paper crane unfolds on a desk, cinematic light"The image command defaults to gpt-image-2-vip at 1K and low quality. To try its other quality levels, add --image-quality medium or --image-quality high. To use the standard model, add --image-model gpt-image-2; that model uses its fixed medium quality, so the script omits the quality field. See the credits and billing steps below before making live calls.
The video command uses seedance-2 with a 4-second duration, 720p resolution, 16:9 aspect ratio, and generate_audio=false. Review Seedance 2 parameters and pricing before making a live video call.
Image and video requests return a task ID first. The script checks its status every 8 seconds, for up to 15 minutes, and prints the result URL when complete. Generated links may expire; save output you need promptly. Use --interval and --max-wait to change polling. If your terminal times out after submission, keep the printed task ID and query the image task or video task endpoint directly.
This is a single-request quickstart, not a batch runner. For production use, add bounded concurrency, persistent task IDs, result storage, and a workload budget in your application. Check account limits and model-specific parameters before increasing volume. After an interrupted generation request, inspect the existing task before resubmitting to avoid duplicate work and charges.
For integration planning, request 1:1 support with your use case, models, estimated monthly usage, peak concurrency, and launch timeline. Capacity and dedicated support scope are assessed for the project.
For API troubleshooting, send the endpoint, model, HTTP status, task/request ID, and redacted reproduction. Use repository Issues for example-code defects; contact support privately for account and billing questions. Never share API keys.
New accounts receive 10 credits—enough to generate 26 images with gpt-image-2-vip at 1K resolution and low quality under current pricing. Higher quality settings or other models use different amounts of credits. The 10 signup credits do not cover this repository's Seedance video example; review pricing and top up before making a live video call. Public price displays can lag updates. Check the current price list before changing models or settings. Actual credit usage is determined by your account's Usage Logs.
After a live call, copy the task ID printed by the script. Sign in to the Dashboard with the same account. Open Task Logs to find the image or video task and its result, and use Usage Logs to inspect the charge or credit usage for the request. If the entry is delayed, refresh the logs and compare the account balance after it appears. The amount recorded in the account is the final charge.
When the free credits are used, open Billing / Top up in the Dashboard to review available recharge options. Review the amount and payment details there before paying.
POST /v1/chat/completions
POST /v1/images/generations
GET /v1/images/generations/{task_id}
POST /v1/videos/generations
GET /v1/videos/generations/{task_id}
The default base is https://toapis.com. See the API documentation for parameters, response formats, and other models. Browse the model catalog before changing the model IDs in the script.
| Symptom | First check |
|---|---|
TOAPIS_API_KEY error |
Set the variable in the same terminal session. |
| HTTP 401 | Check that the key is active and copied correctly. |
| HTTP 400 | Check model availability and request parameters in the current docs. |
| HTTP 429 | Check your account limits and retry later. |
| Task times out | Use the printed task ID to inspect status; video can take longer than the default wait. |
python -m unittest discover -s testsThe automated tests check request shapes and do not make paid calls. Manual live smoke checks completed on 2026-10-09 for the default chat, image, and video workflows. Issues and pull requests that keep the examples aligned with the official docs are welcome.
Documentation checked: 2026-10-09. License: MIT.
Browse 10 real API asset tutorials and cost notes
Browse model types and example costs for 25 tutorials
8 models, one prompt: outputs ordered by example amount. gpt-image-2(VIP) low/1K is $0.0019/image. Cost evidence and project validation scope are stated separately.