Side-by-side comparison
9Router vs LiteLLM: which self-hosted gateway fits your stack?
9Router and LiteLLM are both self-hosted gateway options with different centers of gravity. 9Router is oriented toward coding-tool routing across provider tiers; LiteLLM is positioned as an AI gateway and SDK with unified provider access, spend controls, virtual keys, observability, and enterprise governance.
9Router
LiteLLMQuick answer
Start from the job you need done — the winner is the fit, not the brand.
Choose 9Router if…
Developers who want one local endpoint for Claude Code, Codex, Cursor, Cline, Copilot, OpenCode, and other compatible tools, with provider fallback and a dashboard instead of manual endpoint switching.
Starting point: Free and open source under the MIT license; self-hosted through the npm package or source.
Choose LiteLLM if…
Developers and platform teams who want a self-hosted, free AI gateway with granular spend control, virtual keys, and fallbacks — without per-token platform fees.
Starting point: Open source and self-hosted is free forever ($0): 100+ providers behind one OpenAI-compatible API with...
Comparison table
The same fields are shown for both products so the comparison does not quietly favor one side.
| Decision | 9Router | LiteLLM |
|---|---|---|
| Best for | Developers who want one local endpoint for Claude Code, Codex, Cursor, Cline, Copilot, OpenCode, and other compatible tools, with provider fallback and a dashboard instead of manual endpoint switching. | Developers and platform teams who want a self-hosted, free AI gateway with granular spend control, virtual keys, and fallbacks — without per-token platform fees. |
| Core uses | CLI/IDE model routing, provider fallback, quota tracking, token reduction | AI gateway / LLM proxy, model routing and fallbacks, spend tracking and budgets, virtual API keys and teams |
| Pricing note | Free and open source under the MIT license; self-hosted through the npm package or source. The gateway has no subscription price, but connected providers, accounts, quotas, and their Terms of Service still determine the real cost and availability. Verify live limits before relying on a provider. | Open source and self-hosted is free forever ($0): 100+ providers behind one OpenAI-compatible API with virtual keys, users and teams, spend tracking, budgets and rate limits, LLM fallbacks, request/response logging, and Prometheus metrics. Enterprise is priced annually by gateway request capacity (never per token) and adds SSO/SCIM, OIDC/JWT auth, audit logs, secret managers and key rotation, org and team admins, multi-region control plane, 24/7 support with SLAs, and air-gap deployment; a 30-day trial key is available without a credit card. from litellm.ai/pricing. |
| Free plan | Yes — The gateway software is free and open source (MIT) when self-hosted. | Free tier available |
| Main checks | Self-hosted: you install, secure, update, and monitor the gateway yourself. Provider access is not created by 9Router; subscription, OAuth, API-key, regional, quota, and provider-policy limits still apply. The project documents proxy and MITM-style integrations for some IDE subscriptions; review each provider and client policy before using those routes. Provider counts, free-tier availability, token-savings claims, and release behavior can change; the official site and repository should be rechecked before a purchase or rollout. | Self-hosted: you run and maintain the gateway (PostgreSQL plus optional Redis), so uptime, upgrades, and operations are on you. Enterprise governance features (SSO/SCIM, audit logs, secret managers, SLAs) require the paid Enterprise edition. Independent open-source project (BerriAI) — not an official product of OpenAI, Anthropic, Google, or any listed provider. Security: on March 24, 2026, PyPI releases 1.82.7 and 1.82.8 of the litellm package were published with credential-stealing malware (live about 40 minutes; roughly 2,500 organizations potentially exposed). Pin versions, verify package checksums, and follow the official advisory (BerriAI/litellm issue #24518) before self-hosted installs. Fast release cadence: v1.105.0 (Oct 11, 2026) added day-one Claude Sonnet 5.5 and GPT-6.1 Sol support, a Decisions API and litellm.agent(); v1.104.0 (Oct 3, 2026) added Claude Opus 5.5 and GPT-6. The release notes say to review breaking changes before upgrading, so pin your version and read them first (docs.litellm.ai/release_notes, checked Oct 13, 2026). |
| Pricing source | 9router.com · checked | litellm.ai · checked |
The short answer
Test 9Router when the main problem is connecting coding CLIs and IDEs to a local endpoint with fallback and quota visibility.
Test LiteLLM when the main problem is serving applications or teams with virtual keys, budgets, spend tracking, rate limits, logging, load balancing, and a documented gateway/SDK surface.
Scope and abstraction
9Router emphasizes coding-tool integrations, local OpenAI-compatible access, provider connections, quota tracking, format translation, and three-tier fallback.
LiteLLM’s official docs describe a unified interface for 100+ LLMs, an OpenAI-format gateway, routing and fallback logic, virtual keys, cost tracking, and an admin UI.
Operations and governance
9Router requires the operator to own deployment, credentials, dashboard exposure, updates, and policy review for each route. Its token-saving modes should be benchmarked on your prompts.
LiteLLM also requires deployment and maintenance, while its platform-oriented surface foregrounds spend, identity, teams, rate limits, observability, and load balancing. Confirm open-source versus Enterprise boundaries.
Provider access is not interchangeable
Neither product grants provider access. Check authentication, model support, context, streaming, tool calls, regional availability, retention, and upstream terms for the exact route.
Do not use provider-count claims as a quality score. Compatibility with your client and governance needs matters more than catalog size.
A practical bake-off
Use one synthetic application and the same providers. Compare setup, API compatibility, fallback, spend visibility, authentication, logs, latency, and key rotation. Choose 9Router for coding-tool routing simplicity or LiteLLM for platform controls when the added operations are justified.
Before choosing a paid plan
Match the workflow first. Check current credits, export limits, cancellation terms and commercial-use rights before purchasing.
9Router
Best for: Developers who want one local endpoint for Claude Code, Codex, Cursor, Cline, Copilot, OpenCode, and other compatible tools, with provider fallback and a dashboard instead of manual endpoint switching.
Pricing: Free and open source under the MIT license; self-hosted through the npm package or source. The gateway has no subscription price, but connected providers, accounts, quotas, and their Terms of Service still determine the real cost and availability. Verify live limits before relying on a provider.
Check before paying: Self-hosted: you install, secure, update, and monitor the gateway yourself. Provider access is not created by 9Router; subscription, OAuth, API-key, regional, quota, and provider-policy limits still apply.
LiteLLM
Best for: Developers and platform teams who want a self-hosted, free AI gateway with granular spend control, virtual keys, and fallbacks — without per-token platform fees.
Pricing: Open source and self-hosted is free forever ($0): 100+ providers behind one OpenAI-compatible API with virtual keys, users and teams, spend tracking, budgets and rate limits, LLM fallbacks, request/response logging, and Prometheus metrics. Enterprise is priced annually by gateway request capacity (never per token) and adds SSO/SCIM, OIDC/JWT auth, audit logs, secret managers and key rotation, org and team admins, multi-region control plane, 24/7 support with SLAs, and air-gap deployment; a 30-day trial key is available without a credit card. from litellm.ai/pricing.
Check before paying: Self-hosted: you run and maintain the gateway (PostgreSQL plus optional Redis), so uptime, upgrades, and operations are on you. Enterprise governance features (SSO/SCIM, audit logs, secret managers, SLAs) require the paid Enterprise edition.
Run it 24/7 on your own server
9Router and LiteLLM run on your own machine, so it stops answering when your laptop sleeps. A small Linux VPS with Docker keeps it online for scripts, agents and teammates, with a fixed monthly cost instead of per-token markup.
- Start with the smallest plan that runs Docker; upgrade only if you add users or local models.
- Keep the gateway bound to localhost or behind a firewall, and store provider keys as environment variables.
- Monthly billing lets you test before committing to a long term.
Some links are affiliate links. We may earn a commission if you buy through them. Payment never changes our rankings or editorial verdicts.
Questions creators ask
- Is LiteLLM only for enterprises?
- No. Its official materials describe an open-source library and self-hosted proxy, while Enterprise adds governance, security, and support features. Verify the current edition boundaries.
- Which one gives free access to models?
- Neither guarantees free model access. Provider or local-runtime pricing, quotas, authentication, and policy still apply.
Continue the decision
Use the source-cited profiles and workflow guide before selecting a paid plan.
Sources
Verify changing features, pricing, and usage rights before purchasing.