VernLLM (@VernLLM) header
VernLLM (@VernLLM) profile picture

VernLLM

@VernLLM

The LLM call framework | Resilience, observability and control for every call Managed by @BuddoYap

vernllm.dev
Joined October 2026
  • Tweets26
  • Following1
  • Followers19
9Media
Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)Photo by VernLLM (@VernLLM)
Your LLM call fails behind a gateway and all you get back is a 502. Was it the AI provider? A 429? A 200 that said "N/A"? VernLLM runs in your app, so you see the real error, and keeps the call working: it retries, falls back in order and catches 200s that lie. npm i vern-llm
2
2
1
3
574
Purple, Red, Green, Pink, Yellow and Brown.
3
𝗔dapters, 𝘐ntegrations and E̲xtensions.
1
2
116
VernLLM retweeted
Buddo (@BuddoYap) profile picture
Buddo@BuddoYapOct 14
Building with AI? Before paying for a gateway, check out vern-llm. Genuinely advanced circuit breaking, rate limits, and multi-provider fallback, all in your own process. vernllm.dev
VernLLM: Resilience, observability, and control for every call.

VernLLM: Resilience, observability, and control for every call.

The LLM call framework. Retry budgets, provider fallback, rate limiting and circuit breaking for every call, typed and dependency-light.

vernllm.dev
3
1
1
38
VernLLM retweeted
VernLLM (@VernLLM) profile picture
VernLLM@VernLLMOct 11
Your LLM call fails behind a gateway and all you get back is a 502. Was it the AI provider? A 429? A 200 that said "N/A"? VernLLM runs in your app, so you see the real error, and keeps the call working: it retries, falls back in order and catches 200s that lie. npm i vern-llm
2
2
1
3
574
Why in process beats a gateway, and how VernLLM compares with the AI SDK, LiteLLM and Portkey, feature by feature: vernllm.dev/docs/why

Why VernLLM

How VernLLM compares with the Vercel AI SDK and AI Gateway, LiteLLM, and Portkey on resilience, observability, and control

vernllm.dev
2
38
Hit your AI provider's rate limit and your retries keep you stuck there? The retry budget in VernLLM fixes that. Use it today with "npm i vern-llm".
1
1
2
92
vernllm.dev/docs/core/retry-…
Retry Budget

Retry Budget

Cap how much of a provider's traffic can be retries with a VernLLM retry budget, so a provider outage does not turn into a retry storm.

vernllm.dev
1
38
VernLLM reached 10k+ npm downloads. Advanced logging and routing middleware for LLMs is coming soon.
1
1
56
VernLLM retweeted
Buddo (@BuddoYap) profile picture
Buddo@BuddoYapOct 7
VernLLM is improving more on AEO now
Image from Buddo (@BuddoYap)
Image from Buddo (@BuddoYap)
1
2
139
VernLLM reached 10k+ downloads on npm!
Image from VernLLM (@VernLLM)
1
1
1
3
197
VernLLM retweeted
VernLLM (@VernLLM) profile picture
VernLLM@VernLLMOct 5
Normal circuit breakers trip on any error. For LLMs that's wrong: bad JSON or a caller's 400 isn't an outage. VernLLM's breaker counts only real provider failures, catches empty 200s and stalled streams, and falls back across providers. Use it today with "npm i vern-llm".
4
4
5
96
Normal circuit breakers trip on any error. For LLMs that's wrong: bad JSON or a caller's 400 isn't an outage. VernLLM's breaker counts only real provider failures, catches empty 200s and stalled streams, and falls back across providers. Use it today with "npm i vern-llm".
4
4
5
96
vernllm.dev/docs/core/circui…
Circuit Breaker (Core)

Circuit Breaker (Core)

Stop sending requests to an LLM provider that is down. Set the VernLLM circuit breaker threshold and cooldown, and learn how half open recovery works.

vernllm.dev
26
VernLLM vs AI SDK + AI Gateaway, LiteLLM + Proxy and Portkey + Gateaway in observability
Image from VernLLM (@VernLLM)
9
1
8
225
vernllm.dev/docs/why
Why VernLLM

Why VernLLM

How VernLLM compares with the Vercel AI SDK and AI Gateway, LiteLLM, and Portkey on resilience, observability, and control

vernllm.dev
69
VernLLM retweeted
Buddo (@BuddoYap) profile picture
Buddo@BuddoYapOct 4
RPM + TPM + concurrency with adaptive AIMD is rare, even for a resilient rate-limiter package.
VernLLM (@VernLLM) profile pictureVernLLM@VernLLMOct 4
Rate-limiting LLM calls is a pain. Why not try "npm i vern-llm" and fix it?
2
1
2
60
Rate-limiting LLM calls is a pain. Why not try "npm i vern-llm" and fix it?
1
1
1
2
140
vernllm.dev/docs/core/rate-l…
Rate Limiting (Core)

Rate Limiting (Core)

Stay under a provider's requests/tokens/concurrency limits before they reject you

vernllm.dev
45
VernLLM retweeted
VernLLM (@VernLLM) profile picture
VernLLM@VernLLMOct 1
Your LLM will go down. Your app doesn't have to. VernLLM retries the failing provider, falls back to the next, and after 3 failures opens the circuit so later calls skip it in 0ms. pnpm add vern-llm vernllm.dev How do you handle provider outages today?
9
2
1
11
156
Load more