Is BlockRun a hosted LiteLLM?

Close in shape, different in who holds the accounts. Both give you one OpenAI-compatible API in front of many models. LiteLLM is software you run with your own provider keys; BlockRun is a hosted endpoint that holds the provider relationships and pays them for you.

Do I need my own provider API keys with BlockRun, as I do with LiteLLM?

No. With LiteLLM you configure each provider and manage its key and balance. BlockRun settles with the providers itself, so you call one endpoint and never open an account with any of them.

Can I self-host BlockRun the way I self-host LiteLLM?

The gateway is hosted. The router in front of it, ClawRouter, is open source and runs on your own machine. If the gateway itself must run inside your network, LiteLLM is built for that.

Which costs less, LiteLLM or BlockRun?

LiteLLM's open-source proxy is free, and you pay each provider plus the cost of running the proxy and its database. BlockRun charges each provider's own token rate plus a flat per-call fee, with nothing to run.

Does BlockRun have virtual keys and budgets like the LiteLLM proxy?

Differently. Each pay-per-call request carries its own payment, so an agent can only spend what it signs for that call. On the API-key door, usage is metered per key and billed monthly. LiteLLM's per-team virtual keys and budgets are a deeper internal-governance tool.

LiteLLM vs BlockRun for an autonomous agent: which one?

BlockRun, if the agent has to pay for its own calls with no human holding provider accounts. LiteLLM, if a platform team already owns those accounts and wants to govern how internal apps use them.

Can LiteLLM route to BlockRun as a provider?

Yes. Add BlockRun as an OpenAI-compatible provider in the LiteLLM proxy config and its models appear alongside the rest, with BlockRun paying the upstream providers.

Does BlockRun cover more than the chat models LiteLLM proxies?

Yes. The same endpoint serves image, video, music and speech generation, web search, market data and multi-chain RPC, each priced per call before it runs.

Comparison

BlockRun vs LiteLLM

Updated October 25, 2026

LiteLLM is an open-source proxy and Python SDK that you run yourself. It puts the providers you already hold keys with behind one OpenAI-format API, with virtual keys, budgets and logging for your teams. BlockRun is hosted and holds the provider relationships itself, so a call needs no provider account at all.

	BlockRun	LiteLLM
Deployment	Hosted endpoint; nothing to run. ClawRouter, the router in front of it, is open source and runs on your machine	Self-hosted proxy on your own infrastructure, with your own Postgres and Redis; air-gapped deployments supported
Provider accounts	None needed. BlockRun pays each provider for you	You configure the providers, and the keys and credit with each are yours to manage
Billing	Two doors: pay per call with no account, or an API key and one monthly invoice by card or invoice	The open-source proxy is free; you pay each provider directly. Enterprise license priced on request
Spend controls	Every call is priced before it runs and paid by the request itself; an agent cannot overspend what it signs	Virtual keys with budgets and rate limits per key, user and team
Beyond chat	Image, video, music and speech generation, web search, market data and multi-chain RPC on the same endpoint	Whatever the providers you configure serve
Open source	ClawRouter (MIT)	The gateway (MIT)
Who runs it	BlockRun operates the gateway, status page and upstream relationships	Your team operates the proxy, its database and its upgrades

When to use which

Use LiteLLM if you already hold provider contracts or credits, you must run the gateway inside your own network, or you want virtual keys, budgets and guardrail integrations across many internal teams. It is a mature open-source project and a sensible default for a platform team.

Use BlockRun if you would rather not open an account with every provider, run a proxy and its database, or hand an agent a long-lived key. One hosted endpoint covers the models and the image, video, search and data calls around them, priced per call before it runs.

They also stack: LiteLLM can call BlockRun as an OpenAI-compatible provider, so a platform team keeps its virtual keys and budgets while BlockRun holds the provider accounts behind them.

If something on this page is wrong, email hello@blockrun.ai.

Get a BlockRun API key See every model and price

BlockRun vs LiteLLM questions: self-hosting, keys, cost, agents

Is BlockRun a hosted LiteLLM?: Close in shape, different in who holds the accounts. Both give you one OpenAI-compatible API in front of many models. LiteLLM is software you run with your own provider keys; BlockRun is a hosted endpoint that holds the provider relationships and pays them for you.
Do I need my own provider API keys with BlockRun, as I do with LiteLLM?: No. With LiteLLM you configure each provider and manage its key and balance. BlockRun settles with the providers itself, so you call one endpoint and never open an account with any of them.
Can I self-host BlockRun the way I self-host LiteLLM?: The gateway is hosted. The router in front of it, ClawRouter, is open source and runs on your own machine. If the gateway itself must run inside your network, LiteLLM is built for that.
Which costs less, LiteLLM or BlockRun?: LiteLLM's open-source proxy is free, and you pay each provider plus the cost of running the proxy and its database. BlockRun charges each provider's own token rate plus a flat per-call fee, with nothing to run.
Does BlockRun have virtual keys and budgets like the LiteLLM proxy?: Differently. Each pay-per-call request carries its own payment, so an agent can only spend what it signs for that call. On the API-key door, usage is metered per key and billed monthly. LiteLLM's per-team virtual keys and budgets are a deeper internal-governance tool.
LiteLLM vs BlockRun for an autonomous agent: which one?: BlockRun, if the agent has to pay for its own calls with no human holding provider accounts. LiteLLM, if a platform team already owns those accounts and wants to govern how internal apps use them.
Can LiteLLM route to BlockRun as a provider?: Yes. Add BlockRun as an OpenAI-compatible provider in the LiteLLM proxy config and its models appear alongside the rest, with BlockRun paying the upstream providers.
Does BlockRun cover more than the chat models LiteLLM proxies?: Yes. The same endpoint serves image, video, music and speech generation, web search, market data and multi-chain RPC, each priced per call before it runs.