Help Center

This page explains on-site payment, redemption-code purchases, redemptions, API-key creation, and configuration for common clients.

Contents

• 1. Purchasing and payment
• 2. Redeem a code
• 3. Create an API key
• 4. Plans, groups, and API types
• 5. OpenAI-compatible API
• 6. Anthropic / Claude-compatible API
• 7. Claude Code
• 8. Codex
• 9. CC-Switch
• 10. Frequently asked questions
• 11. Security recommendations
• 12. Contact support
• 13. xAI / Grok-compatible API

1. Purchasing and payment

FluxionAPI supports on-site payment and redemption codes from the code store. Use on-site payment first. If it is unavailable or the code store is running a promotion, purchase a code and redeem it in the console.

On-site payment:

1. Sign in.
2. Open “Subscriptions” or “Purchase plan.”
3. Select the desired product.
4. Complete payment as instructed.
5. Confirm that the balance, plan, or entitlement activates automatically.

Redemption-code store:

Purchase a redemption code

Code-store purchase flow:

1. Open the store and select the correct product.
2. Confirm its name, price, promotion, and delivery contents.
3. Complete payment.
4. Copy the automatically delivered code.
5. Return to the FluxionAPI “Redemption code” page and redeem it.

Keep the delivered code secure. If it is not delivered after payment, check the store order first. If a valid code cannot be redeemed, contact support with the order information and code.

2. Redeem a code

A code may activate a plan, add balance, or grant another entitlement according to its product description.

1. Sign in.
2. Open “Redemption code” from the sidebar.
3. Paste the complete code.
4. Select Redeem.
5. Confirm that the entitlement is active.

A code is normally single-use. Avoid leading or trailing whitespace. For an invalid, used, or expired code, verify the purchase record before contacting support.

3. Create an API key

An API key authenticates requests to FluxionAPI.

1. Open “API Keys.”
2. Select “Create API key.”
3. Enter an identifiable name.
4. Choose an available group.
5. Create the key and store it securely.

The complete value may be shown only once. Never disclose it publicly. If exposure is suspected, delete the old key and create a replacement. Different keys can use different groups, and each group determines available models and whether the request uses an Anthropic- or OpenAI-compatible format.

4. Plans, groups, and API types

Plans, groups, models, and protocols are linked. Choose the group matching the active plan and use that group's required API format.

4.1 Plans and groups

Plan	Recommended group	Main models	Standard API type
GPT plan	GPT group	GPT / OpenAI family	OpenAI-compatible
Claude plan	Claude group	Claude family	Anthropic / Claude-compatible
All-model plan	All-model group	GPT and Claude	Follow the group description; commonly Anthropic-compatible

The wrong group can produce an unsupported-model error, no available account, a protocol mismatch, or a request that authenticates but cannot reach the model.

4.2 Groups and protocols

GPT group

Base URL: https://fluxionai.space/v1
Authentication: Authorization: Bearer YOUR_API_KEY
Common endpoint: /v1/chat/completions
Example models: gpt-5.5, gpt-5.4

Claude group

Base URL: https://fluxionai.space
Authentication: x-api-key: YOUR_API_KEY
Common endpoint: /v1/messages
Example models: claude-sonnet-4-6, claude-opus-4-6

All-model group

Follow the purchased plan's group instructions. Use an Anthropic-compatible request for Claude models. For GPT, use the protocol specified by the group or current announcement. If the all-model key is configured as an Anthropic-compatible group, use that format even when requesting a supported GPT model.

4.3 Model names

Model names are character-sensitive. Copy the complete call name from Model Marketplace or the vendor card under Channel Status. Examples on this page do not guarantee current availability; the plan, group, and key permissions are authoritative.

5. OpenAI-compatible API

Base URL: https://fluxionai.space/v1
API Key: your FluxionAPI user key
Model: a GPT model supported by the group

Chat Completions

curl https://fluxionai.space/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $FLUXION_API_KEY" \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      {"role": "user", "content": "Please briefly introduce yourself."}
    ]
  }'

Images / gpt-image-2

Image models do not use Chat Completions. Use the Images endpoint:

Base URL: https://fluxionai.space/v1
Endpoint: /v1/images/generations
Authentication: Authorization: Bearer YOUR_API_KEY
Model: gpt-image-2

curl https://fluxionai.space/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $FLUXION_API_KEY" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "An orange cat beside a window in a cyberpunk city",
    "size": "1024x1024",
    "n": 1,
    "response_format": "b64_json"
  }'

Select an Images or Image Generation client type, not a chat-model type. Channel monitoring may primarily test text endpoints, so its result may not represent image-model availability; verify /v1/images/generations directly.

Enter https://fluxionai.space/v1 when the client asks for a Base URL. Do not create a duplicated path such as https://fluxionai.space/v1/v1.

6. Anthropic / Claude-compatible API

Base URL: https://fluxionai.space
API Key: your FluxionAPI user key
Model: a Claude model supported by the group

Messages

curl https://fluxionai.space/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: $FLUXION_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Please briefly introduce yourself."}
    ]
  }'

OpenAI-compatible calls normally use Authorization: Bearer API_KEY. Anthropic-compatible calls use x-api-key: API_KEY. Client field names can differ, but Base URL, API key, and model remain the core values.

7. Claude Code

Configure Claude Code with:

ANTHROPIC_BASE_URL=https://fluxionai.space
ANTHROPIC_API_KEY=YOUR_API_KEY

Select a Claude model supported by the plan, such as claude-sonnet-4-6.

If a call fails, check that the key is correct, its group supports Claude, the plan is active, sufficient balance or allowance remains, and the Base URL is https://fluxionai.space without /v1.

8. Codex

Codex sends OpenAI Responses requests. GPT and all-model groups use different Base URLs.

8.1 GPT group

Base URL: https://fluxionai.space/v1
API Key: your GPT-group user key
Model: a GPT model supported by the group

Example names include gpt-5.5, gpt-5.4, and gpt-5.4-mini. Let Codex append /responses; do not enter a duplicate /v1. Do not use Claude Code's ANTHROPIC_BASE_URL setting for a GPT group.

8.2 All-model group

Base URL: https://fluxionai.space/vip/v1
API Key: your all-model user key
Model: a model supported by the group

Codex ultimately requests:

https://fluxionai.space/vip/v1/responses

The route converts the Responses request and passes it to FluxionAPI for authentication, billing, group permission checks, fallback/account selection, and upstream forwarding. Do not append /v1/responses manually and do not create a separate proxy key. Long context, upstream availability, and model response time may affect latency without indicating a key error.

8.3 Common Codex errors

• Wrong Base URL: use /v1 for GPT groups and /vip/v1 for all-model groups.
• Wrong key group: a group and endpoint mismatch can cause unsupported-model, no-account, or request errors.
• Wrong model: copy the complete supported name from Channel Status.

9. CC-Switch

CC-Switch switches providers and protocols for Claude Code and Codex. Confirm the client, key group, model, and protocol first.

For current CC-Switch releases, enable local routing whenever Claude Code or Codex uses a GPT model through an OpenAI-compatible or Responses endpoint. If an error occurs, toggle local routing off and on. Under Settings → Routing → Local router, enable the visible router switch, master switch, target client, and the required rectifier features.

9.1 Before configuration

Confirm that:

1. The FluxionAPI account is signed in.
2. A valid plan or entitlement has been purchased or redeemed.
3. A user API key exists.
4. Its group supports the target model.
5. The complete current model name has been copied.
6. The target client is known.

9.2 Claude Code with Claude

Add an Anthropic / Claude-compatible provider:

Name: FluxionAPI Claude Code - Claude
Base URL: https://fluxionai.space
API Key: your user API key
Model: claude-sonnet-4-6
API type: Anthropic / Claude / Claude Code Compatible

Environment-variable form:

ANTHROPIC_BASE_URL=https://fluxionai.space
ANTHROPIC_API_KEY=YOUR_API_KEY
ANTHROPIC_MODEL=claude-sonnet-4-6

The Base URL must not end in /v1.

9.3 Claude Code with GPT

Enable local routing for Claude Code first. Then choose either:

OpenAI-compatible format

Name: FluxionAPI Claude Code - GPT OpenAI
Base URL: https://fluxionai.space/v1
API Key: your user API key
Model: gpt-5.5
API type: OpenAI Compatible

Codex / Responses format

Name: FluxionAPI Claude Code - GPT Codex
Base URL: https://fluxionai.space/v1
API Key: your user API key
Model: gpt-5.5
API type: Codex / OpenAI Responses Compatible

Do not append /responses unless the Full URL workflow requires it. The key's group must support the selected GPT model.

9.4 Codex with GPT

Name: FluxionAPI Codex - GPT
Base URL: https://fluxionai.space/v1
Full URL when explicitly enabled: https://fluxionai.space/v1/responses
API Key: your user API key
Model: gpt-5.5

Use a GPT-group key. When Full URL is disabled, do not append /responses. When it is enabled, use the complete endpoint together with local routing.

9.5 Activate and test

Save the provider, select “Use,” and restart the target client. Send a simple request such as:

Reply with: FluxionAPI configuration succeeded

9.6 CC-Switch errors

• Local router disabled: Claude Code with GPT may bypass the provider or use an incompatible protocol.
• Wrong API type: use Anthropic for Claude; use OpenAI-compatible or Responses for GPT.
• Wrong Base URL: Anthropic uses https://fluxionai.space; OpenAI and Responses commonly use https://fluxionai.space/v1.
• Wrong key group: verify that the selected group supports the model.
• Wrong model name: copy the exact name, including case, hyphens, and version.

10. Frequently asked questions

10.1 Why is no upstream account available?

The model may have no currently available account, may be outside the key's group, may have reached an upstream concurrency limit, or may be under maintenance. Verify the name and group, try another model included by the plan, and retry later.

10.2 Why is the model unsupported?

The group may not include it, the name may be wrong, or the plan may not cover it. Check the key group, plan description, and spelling.

10.3 Why did authentication fail?

Verify the key and header format. OpenAI-compatible APIs use Authorization: Bearer YOUR_API_KEY; Anthropic uses x-api-key: YOUR_API_KEY. Remove accidental spaces and line breaks.

10.4 Why was balance or allowance charged?

Charges are calculated from token usage, model price, and plan rules. Review the request time, model, tokens, and charge under Usage Records.

10.5 What if on-site payment fails?

Refresh the order status first. If payment still cannot be completed, use the redemption-code store.

10.6 What if the purchased code was not delivered?

Check the code-store order status. If payment succeeded without delivery, contact support with the order number, payment time, and product.

10.7 What if a code cannot be redeemed?

Confirm that it is complete, contains no spaces or line breaks, has not been used, has not expired, and applies to the current account and service. Contact support if it remains unusable.

11. Security recommendations

• Never share an API key publicly.
• Never commit one to a public repository.
• Keep keys out of screenshots, videos, and logs.
• Review usage records regularly.
• Revoke a key and contact support after abnormal usage.
• Use separate keys for separate applications so each can be revoked independently.

12. Contact support

Support can help with redemption, API keys, failed model calls, subscriptions, billing, login, and account security. Include the account email, incident time, model, error text or screenshot, and relevant order information.

Email: service@fluxionai.world

For privacy-related inquiries, see Contact Us in the Privacy Policy.

Never send a complete API key in a public channel.

13. xAI / Grok-compatible API

Grok uses the OpenAI-compatible API:

Base URL: https://fluxionai.space/v1
API Key: your user API key
Model: a Grok model supported by the group

Example names include grok-4.5 and grok-4.3.

curl https://fluxionai.space/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $FLUXION_API_KEY" \
  -d '{
    "model": "grok-4.5",
    "messages": [
      {"role": "user", "content": "Please briefly introduce yourself."}
    ]
  }'

Use /v1/images/generations for image models such as grok-imagine, following the image example in section 5.