Cursor on Seedrise

What works in Cursor

Checked against Cursor's docs and release notes on 2 November 2026. Cursor changes often; the sources are listed at the end of this page.

Why run Cursor through Seedrise

Setup in five minutes

1. Check your Cursor plan. Own API keys are not available on Cursor's Plus plan. On Teams and Enterprise, an admin can switch personal API keys off for the team.
2. Create a Seedrise key for Cursor. In the Seedrise console, create a key limited to the models you will add, with a monthly budget. Leave its IP allow-list empty: Cursor sends requests from its own servers, not from your machine.
3. Add the key and base URL. Open Cursor Settings > Models. In the OpenAI section, paste your Seedrise key into API Key and turn on Use API Key. Turn on Base URL override and enter the URL below, ending in /v1 with nothing after it.
   Base URLCopy

   https://api.seedrise.ai/v1

4. Add Seedrise model IDs. Add each model to Cursor's model list by its Seedrise ID, with no vendor prefix, and turn it on. Start with models Cursor does not route to its own providers.
   Model IDsCopy

   deepseek-v4-pro
   kimi-k3.1-pro
   qwen3.8-max
   mistral-large-4
   deepseek-v3.2

5. Pick the model and verify. Choose the model in Agent or Chat and send a prompt. The request should appear in the Seedrise console within a minute, under the Cursor key.

Models to start with in Cursor

Provider list prices in USD per 1M tokens, input / output, as at 1 November 2026. Context windows checked 1 November 2026. Your Seedrise rate card is quoted per account against these prices.

Models marked AU can be pinned to Sydney or Melbourne on request. Every other ID is in the model library; see plans and pricing for how rates are set.

Check it is working

• The Seedrise console shows the request under the Cursor key and the model you picked. If nothing appears, the request did not reach Seedrise.
• To rule out the key, call Seedrise directly. The first command lists the IDs your key can call; the second sends one message.

curl https://api.seedrise.ai/v1/models -H "Authorization: Bearer sk-..."

curl https://api.seedrise.ai/v1/chat/completions \
  -H "Authorization: Bearer sk-..." \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-pro", "messages": [{"role": "user", "content": "Hello"}]}'

• Tab suggestions keep coming from Cursor's own models and never show in the Seedrise console.

Known errors

Still stuck? Email support@seedrise.ai with the time of the request, the model ID and the key name.

For teams

Each developer adds the Seedrise key in their own Cursor settings. Give each team its own key so spend shows up per project on the invoice.

• Enterprise admins control personal API keys under Team Settings > Models. If the OpenAI key field is unavailable, ask your Cursor admin.
• On Cursor Teams and Enterprise plans, Cursor can still charge its own per-token rate on requests made with your own key. Seedrise bills the model usage; Cursor bills its rate.
• Cursor's Zero Data Retention policy does not apply to requests made with your own keys; data handling follows the key provider, here Seedrise. See security and data.
• Control Cursor keys with budget, model allow-list and expiry. An IP allow-list blocks Cursor's servers.
• For Claude across a team, use Claude Code on Seedrise, where managed settings lock every developer to the gateway.

Cursor questions

Sources

• Supported key fields, chat models only, Tab, routing through Cursor's servers and Zero Data Retention: cursor.com/help/models-and-usage/api-keys, checked 2026-11-02.
• API Key fields, the Use API Key switch and the OpenAI Base URL override (version 3.20): cursor.com/docs/release-notes, checked 2026-11-02.
• Plans with and without own API keys, Cursor's model list and its per-token rate on Teams and Enterprise: cursor.com/docs/models-and-pricing, checked 2026-11-02.
• Admin control of personal API keys: cursor.com/docs/enterprise/model-and-integration-management, checked 2026-11-02.
• Which models use the OpenAI override, Cursor staff reply, 23 September 2026: forum.cursor.com, checked 2026-11-02.