Pricing and Usage

SandBase uses pay-as-you-go billing. Model availability, price formulas, cache multipliers, and capability-specific charges can change independently, so integrations should read live metadata instead of relying on copied price tables.

Daily free API calls

Every organization gets 100 free data API calls (search, social, scraping, e-commerce, finance) per UTC day. The allowance is shared by all eligible data APIs and all of the organization's keys, and resets at 00:00 UTC. LLMs and media models are not eligible. Some data APIs are not eligible either: pricier or formula-priced ones, such as some batch endpoints, crawl endpoints, Xiaohongshu PGY, and the Douyin Xingtu kol-* routes, are billed from the first call. To check a tool, run sandbase_inspect: eligible tools carry pricing.free_quota, and tools without it are always billed.

• Calls within the allowance are not charged, and an organization with a zero balance can still use them.
• After the 100th call, eligible calls keep working and are billed per call at the price the tool shows. They are not rejected or rate-limited because the allowance is used up.
• Unused calls do not carry over to the next day.

In the SandBase MCP tools, sandbase_inspect marks eligible tools with pricing.free_quota (daily_limit, scope, resets). sandbase_account returns today's allowance as free_api_quota with daily_limit, used_today, remaining, and resets_at; used_today and remaining are omitted when the count is temporarily unavailable. After a run settles, sandbase_run and sandbase_run_get report free_quota_status: free (covered by the allowance), overage (billed after the allowance was used up), or degraded (not charged because the allowance count was unavailable).

Manage billing in the Console

Use Developer → Billing to view your balance, buy credits, set up auto top-up, redeem coupons, and manage the billing settings available to your organization. Open Developer → Activities → Usage for organization-level request, token, and cost summaries.

Public API boundary

Balance, top-up, billing export, and organization-wide usage endpoints are not currently part of the supported public API. Do not automate Console network requests or depend on internal Console routes.

Read current model pricing

First discover an enabled model:

curl 'https://api.sandbase.ai/v1/models?q=gpt' \
  -H "Authorization: Bearer $SANDBASE_API_KEY"

Then request its full model card:

curl https://api.sandbase.ai/v1/models/deepseek/deepseek-v4-flash \
  -H "Authorization: Bearer $SANDBASE_API_KEY"

The list endpoint is optimized for discovery and intentionally omits detailed token, cache, and reasoning prices. Read the selected model's detailed pricing from model_card in the detail response. There is no universal formula that applies to every LLM, media model, embedding model, or API capability.

Check one task's recorded cost

When an operation returns a task ID, query its settlement record:

curl https://api.sandbase.ai/v1/tasks/{task_id}/cost \
  -H "Authorization: Bearer $SANDBASE_API_KEY"

The response distinguishes the current cost from estimated_cost and reports whether settlement is final. Do not use cost for billing reconciliation until settled is true. Task records are operational lookup data rather than an accounting ledger. Use the billing records and CSV exports available under Developer → Activities → Usage for reconciliation and longer-term records.

Cost-control practices

• Resolve the current model card before launching a new workload.
• Check capability_tags, supported_modes, and unified_schema; cheaper alternatives are useful only when they support the required input and output behavior.
• Limit unnecessary prompt context and set an appropriate output-token limit when the selected model supports it.
• Treat caching as model-specific. Check the model card and returned usage fields instead of assuming a provider-wide discount.
• Record task IDs only for the task-cost lookup; for asynchronous /v1/run work, record the returned opaque run ID and poll GET /v1/run/{id} only while status is pending or running.
• Use spending controls exposed in the Console for your organization.

Payment errors

An API operation may return HTTP 402 when the organization cannot fund the request. Error envelopes vary by API compatibility surface, so handle the HTTP status and documented error object rather than matching one message string. After resolving billing in the Console, retry according to the operation's idempotency and recovery rules.

Video submissions are also checked against their estimated cost before they start. A single top-up of $10 or more raises your account level, which allows more video tasks to run at the same time. See Concurrency limits.

See Models and Pricing, Task Cost, and Errors for the current public contracts.