Question 1

How do I integrate Fivo?

Accepted Answer

Change one URL in your code. Replace your AI provider's API base URL (e.g., api.openai.com) with your Fivo endpoint. Same API format, same parameters, zero other code changes. Takes under 2 minutes.

Question 2

Which programming languages does Fivo support?

Accepted Answer

Fivo works with any language that can make HTTP requests: Python, TypeScript/JavaScript, Go, Java, Ruby, Rust, C#, PHP, and more. It is a drop-in replacement at the API level, not a language-specific SDK.

Question 3

Does Fivo work with the official OpenAI Python SDK?

Accepted Answer

Yes. Set the base_url parameter in the OpenAI client to your Fivo endpoint. All other code stays exactly the same. Example: client = OpenAI(base_url='https://api.Fivo.dev/v1', api_key='your-openai-key').

Question 4

Does Fivo work with LangChain?

Accepted Answer

Yes. Change the base_url parameter in your ChatOpenAI or OpenAI LangChain configuration. Works with LangChain, LangGraph, and all LangChain ecosystem tools. Zero code changes beyond the URL.

Question 5

Does Fivo work with LlamaIndex?

Accepted Answer

Yes. Set the api_base parameter in your LlamaIndex OpenAI configuration to your Fivo endpoint. Works with all LlamaIndex components including query engines, agents, and RAG pipelines.

Question 6

Which AI providers does Fivo support?

Accepted Answer

Fivo supports 9+ providers: OpenAI (GPT-4o, GPT-4, GPT-3.5 Turbo), Anthropic (Claude 3.5, Claude 3), Google (Gemini Pro, Ultra), DeepSeek, Mistral (Large, Medium), Cohere (Command R+), Together AI, Azure OpenAI, and any OpenAI-compatible endpoint.

Question 7

Do I need to change my API keys?

Accepted Answer

No. You keep your existing AI provider API keys. Fivo uses them to route requests when needed. Your keys are encrypted at rest and never logged. You can also use Fivo-provided keys if preferred.

Question 8

Does Fivo support function calling / tool use?

Accepted Answer

Yes. Full support for OpenAI function calling, Anthropic tool use, and structured outputs. Fivo handles these transparently � your function definitions, tool schemas, and response parsing all work identically.

Question 9

Does Fivo support streaming (SSE)?

Accepted Answer

Yes. Full Server-Sent Events streaming support for all providers. Optimized streaming responses start in under 5ms vs 200-800ms for direct API calls. Your existing streaming code works without changes.

Question 10

Does Fivo work with embedding models?

Accepted Answer

Yes. Fivo supports OpenAI embedding models (text-embedding-3-small, text-embedding-3-large, text-embedding-ada-002) and Cohere embed models. Embedding requests benefit from the same optimization and cost reduction.

Question 11

Can I use Fivo with multiple AI providers simultaneously?

Accepted Answer

Yes. Fivo's smart model selection automatically routes queries to the best provider based on complexity, cost, and your quality requirements. Or you can specify the model explicitly and Fivo handles the provider routing.

Question 12

Does Fivo work with Semantic Kernel?

Accepted Answer

Yes. Change the endpoint URL in your Semantic Kernel OpenAI connector configuration. Works with Semantic Kernel for .NET and Python. All kernel functions, planners, and plugins work identically.

Question 13

Does Fivo work with AutoGen / CrewAI?

Accepted Answer

Yes. Both AutoGen and CrewAI use OpenAI-compatible clients under the hood. Change the base URL in your agent configuration and Fivo optimizes all agent queries automatically.

Question 14

Can I integrate Fivo with my existing API gateway?

Accepted Answer

Yes. Fivo works behind any API gateway (Kong, AWS API Gateway, Cloudflare, etc.). Route your AI API traffic through Fivo before it reaches the provider. Fivo acts as a transparent optimization layer.

Question 15

Is there a REST API for Fivo management?

Accepted Answer

Yes. Fivo provides a management API for programmatic access to analytics, configuration, team management, and billing. Available on Pro and Enterprise plans. Full OpenAPI documentation included.

Question 16

How do I monitor Fivo performance?

Accepted Answer

Fivo provides a real-time dashboard showing cost savings, query volume, optimization rates, latency metrics, and provider usage. Enterprise plans support webhook notifications, Slack alerts, and export to external monitoring systems.

Question 17

Can I roll back from Fivo to direct API calls?

Accepted Answer

Yes. Simply change the URL back to your provider's endpoint. No data migration, no cleanup, no lock-in. Your application works immediately with the direct API. Fivo is designed for zero vendor lock-in.

Question 18

Does Fivo add latency for non-optimized queries?

Accepted Answer

Minimal. Non-optimized queries add approximately 5-15ms of routing overhead. However, Fivo's connection pooling and pre-warming often reduce total latency by 2-3x even for non-optimized queries.

Question 19

How does Fivo handle rate limits?

Accepted Answer

Fivo manages rate limits across providers automatically. If one provider is rate-limited, queries route to alternatives seamlessly. Fivo's optimization also reduces the number of API calls reaching your provider, so you are less likely to hit rate limits.

Question 20

Does Fivo support batch API requests?

Accepted Answer

Yes. Fivo supports batch processing through both synchronous and asynchronous request patterns. Each request in a batch benefits from the same optimization. Use parallel requests for maximum throughput.

Question 21

Can I use Fivo in a serverless environment?

Accepted Answer

Yes. Fivo works in AWS Lambda, Vercel, Cloudflare Workers, Google Cloud Functions, and Azure Functions. Since Fivo is a remote endpoint, there is no local state or cold-start overhead to worry about.

Question 22

Does Fivo support TypeScript / Node.js?

Accepted Answer

Yes. Works with the official OpenAI Node.js SDK, Anthropic SDK, and any HTTP client (axios, fetch, etc.). Set the baseURL configuration option to your Fivo endpoint. Same API, zero code changes.

Question 23

How do I test Fivo before going to production?

Accepted Answer

Use the free Starter plan (1,000 queries/day). Point your staging or development environment to Fivo and monitor the dashboard. Compare cost savings, response times, and accuracy. When satisfied, switch production traffic.

Question 24

Does Fivo work with custom fine-tuned models?

Accepted Answer

Yes. Fivo supports OpenAI fine-tuned models (ft: prefix). Optimizations work the same way for custom models. Specify your fine-tuned model ID in requests as usual.

Question 25

What is the maximum request size Fivo supports?

Accepted Answer

Fivo supports requests up to 128K tokens (matching the largest context window available). There is no practical limit on request size beyond what the underlying AI provider supports.

Framework / Language	What to Change	Time
Python (OpenAI SDK)	`base_url` parameter	30 seconds
TypeScript (OpenAI SDK)	`baseURL` option	30 seconds
LangChain	`openai_api_base`	30 seconds
LlamaIndex	`api_base`	30 seconds
Semantic Kernel	Endpoint URL in connector	1 minute
cURL / REST API	URL in request	10 seconds
Any OpenAI-compatible	API base URL	30 seconds

Fivo Integration FAQ

Setup Overview

All 25 Integration Questions