Valentina
09/17/2025, 5:36 PMVapi Support Bot
09/17/2025, 5:37 PMserver.url in the tool definition) or as built-in Vapi functions? This will help determine if variable passing (like {{get_auth_token.access_token}}) should work natively or if a custom orchestration is needed.
Sources:
- [Custom LLM Tool Calling Integration](https://docs.vapi.ai/customization/tool-calling-integration)
- [Custom Tools](https://docs.vapi.ai/tools/custom-tools)Valentina
09/17/2025, 7:31 PMaccess_token
- **Tool 2**: Built-in API Request → POST /resource with Authorization: Bearer {{get_auth_token.access_token}}
**Issue**: The variable {{get_auth_token.access_token}} doesn't resolve - header arrives empty.
**Also Tried**: Workflows with 2 API Request nodes and {{access_token}} variable passing, but workflow doesn't start conversation (assistant doesn't speak, call ends immediately).
**Questions**:
1. Should {{tool_name.variable}} work between built-in API Request functions?
2. Is there a way to implement OAuth → Business API flow using only Vapi's built-in functions?
3. Do workflows support variable passing between API nodes, or are the audio issues a known limitation?
I'm trying to keep everything within Vapi's native functionality without external servers. Is there a configuration I'm missing for this OAuth pattern?
Thanks!Toketo1
09/17/2025, 7:41 PMValentina
09/17/2025, 7:49 PM{{tool_name.variable}} should work between built-in API Request functions
2. If there's a correct way to configure this OAuth flow within Vapi's native tools
3. Any known issues or limitations with workflows for this use case
If you have documentation links or configuration examples that would be perfect. I understand if some complex issues require more detailed discussion, but any written pointers would be greatly helpful.
Thanks again for the support!Toketo1
09/17/2025, 8:07 PMToketo1
09/17/2025, 8:08 PMToketo1
09/17/2025, 8:08 PMToketo1
09/17/2025, 8:09 PM