# Liquid Inference vs Helicone

Helicone is an observability platform with a gateway. It logs each request, and its gateway sends the request to the cheapest provider that it knows, with your keys or with Helicone credit. Liquid Inference is an exchange. Providers quote their own prices, and each request starts an auction among the offers that meet your rules.

The facts about Helicone come from its own pages, checked on 10 November 2026. The links are in [Sources](#sources).

## Feature by feature

| Feature                   | Helicone                                                                      | Liquid Inference                                                   |
| ------------------------- | ----------------------------------------------------------------------------- | ------------------------------------------------------------------ |
| What you buy              | Logs and a gateway. A plan: free, $79 or $799 a month, or Enterprise          | Inference. One key and one balance for each provider               |
| Price                     | The provider's price                                                          | Each provider quotes its own price                                 |
| Provider choice           | The cheapest first. Providers at equal cost share the load                    | An auction: the lowest charge plus wait cost wins                  |
| Fee                       | The plan. Helicone credit, where open, has no token markup; you pay card fees | 5% of the charge, or a floor on cheap models. No plan fee          |
| Cap for each request      | No. A cost limit for each time window                                         | Yes. `cap_usd`; the provider pays any usage above it               |
| Preview with no charge    | Partly. A public calculator from token counts that you type                   | Yes. The charge and cap of each offer, and why others are out      |
| Record of each request    | Partly. A log with a cost that Helicone estimates                             | Yes. A receipt at the rates of the offer                           |
| No prompt logs by default | No. Kept 7 days to 3 months on public plans                                   | Yes. The exchange stores no prompt or answer                       |
| Open to sellers           | No. Helicone adds the providers                                               | Yes. By application; sellers post their own offers                 |
| API formats               | Partly. OpenAI formats. Anthropic Messages only through an older proxy        | Yes. OpenAI Chat Completions and Responses, and Anthropic Messages |
| Media                     | Partly. Image output for one model                                            | Yes. Image, video, speech, transcription and embeddings            |
| Self-host                 | Yes. Under the Apache 2.0 licence                                             | No                                                                 |

## What the exchange does that Helicone does not

- **Prices from competition.** Each provider quotes its own price for each model, and the auction buys the lowest cost. See [Price and wait](https://liquidinference.ai/docs/price-and-wait.md).
- **A hard cap for each request.** Send `exchange.cap_usd`, or the `x-liquid-cap-usd` header. See [Purchasing rules](https://liquidinference.ai/docs/rules.md).
- **A dry run.** See which offers compete and what the request can cost before you buy. See [Purchasing rules](https://liquidinference.ai/docs/rules.md).
- **An exact charge.** The receipt shows what you paid, from the metered usage and the rates of the offer. See [Receipts](https://liquidinference.ai/docs/receipts.md).
- **No prompt logs.** See [Data handling](https://liquidinference.ai/docs/data-handling.md).

## When Helicone is the better choice

- You want a full log of each prompt and answer, with sessions and scores, for your own analysis.
- You must run the platform in your own network.

## Switch in one line

```diff
 client = OpenAI(
-    base_url="https://ai-gateway.helicone.ai",
+    base_url="https://router.inference.ai.exchange/v1",
     api_key=...,
 )
```

Then change each `model` to a name that `GET /v1/models` lists. See [Connect an app](https://liquidinference.ai/docs/connect-overview.md).

## Questions

### Does the exchange show the cost of each request?

Yes. Each job has a receipt, and the response has an `exchange` block with the price cap and the charge. See [Receipts](https://liquidinference.ai/docs/receipts.md).

### Can the exchange use my own keys?

Yes, in beta. Connect a cloud key, and it competes in the auction. The cloud bills you, and the exchange charges only its fee. See [Bring your own key (BYOK)](https://liquidinference.ai/docs/byok.md).

## Sources

Checked on 10 November 2026.

- Gateway and credit: [Gateway overview](https://docs.helicone.ai/gateway/overview), [Platform overview](https://docs.helicone.ai/getting-started/platform-overview)
- Routing: [Provider routing](https://docs.helicone.ai/gateway/provider-routing)
- Plans and log retention: [Pricing](https://www.helicone.ai/pricing)
- Cost limits: [Custom rate limits](https://docs.helicone.ai/features/advanced-usage/custom-rate-limits)
- Cost record: [How we calculate cost](https://docs.helicone.ai/references/how-we-calculate-cost)
- Calculator: [LLM cost](https://www.helicone.ai/llm-cost)
- Formats and media: [Anthropic integration](https://docs.helicone.ai/getting-started/integration-method/anthropic), [Image generation](https://docs.helicone.ai/gateway/concepts/image-generation)
- Self-hosting: [Helicone on GitHub](https://github.com/Helicone/helicone)