Comments (11)
@davlhd I went through the exact problem. You're probably adding a lot of input context as time goes along (that's the beauty of this thing). I added $100, then $200 and Claude bumps up your tiers. Then I spoke with sales and I'm on t4 now. It only takes a day or so to get unleashed again.
I also highly recommend that you save your chat, review the md file and optimize the context that flows back in. In earlier versions of this repo, it didn't remove files from its input context as it picked up more work, so I think there's definitely some context-management that needs to be added.
@Doriandarko is killing it though with this repo, it's insane. Hope that helps!
from claude-engineer.
You would need OpenRouter
from claude-engineer.
I have OpenRouter working through a new API menu but tools not working well due to request/response differences. Going to look into it more and/or just have it route main and not tools through OR as an option when that endpoint is selected.
Another option, OP, is to set the tools usage to the haiku model, that'll reduce usage quite a bit and haiku seems to work well for tools.
from claude-engineer.
How long do they take to respond for your request to upgrade tiers?
from claude-engineer.
How long do they take to respond for your request to upgrade tiers?
They dont upgrade Tiers. I also asked for it.
from claude-engineer.
Well rip, what do we do then? How are people using this tool, self-hosted? LLAMA 405B?
from claude-engineer.
Anthropic sales/support is extremely lackluster. That being said, usage over time will raise your Tier level; I believe at Tier 3 I got to 5mm requests per day for 3.5; don't think it changed at Tier 4 which I'm now at. So keep using it and at least it'll rise over time.
from claude-engineer.
Oh, so the "solution" is to just max out the usage everyday, go take a break, then come back and repeat tomorrow and it goes up eventually?
from claude-engineer.
Another option, OP, is to set the tools usage to the haiku model, that'll reduce usage quite a bit and haiku seems to work well for tools.
also, haiku apparently has a very low output limit?
Error in tool response: Error code: 400 - {'type': 'error', 'error': {'type': 'invalid_request_error', 'message': 'max_tokens: 8000 > 4096, which is the maximum allowed number of │
│ output tokens for claude-3-haiku-20240307'}}
from claude-engineer.
Oh, so the "solution" is to just max out the usage everyday, go take a break, then come back and repeat tomorrow and it goes up eventually?
Correct, here are the tiers: https://docs.anthropic.com/en/api/rate-limits
They finally got back to me from contacting a couple weeks ago and bumped me to Tier 5 50,0000,000, a 10x increase. I'd min/max usage AND reach out with a plea and what you're working on and see if they don't perhaps kick you up a few tiers. May help to find a few Benjamins in account as well, to show you're serious?
from claude-engineer.
Another option, OP, is to set the tools usage to the haiku model, that'll reduce usage quite a bit and haiku seems to work well for tools.
also, haiku apparently has a very low output limit?
Error in tool response: Error code: 400 - {'type': 'error', 'error': {'type': 'invalid_request_error', 'message': 'max_tokens: 8000 > 4096, which is the maximum allowed number of │
│ output tokens for claude-3-haiku-20240307'}}
Ya you'd need to exit the script and bump tool output tokens to 4096, when using Haiku.
from claude-engineer.
Related Issues (20)
- How to resume from a saved chat? HOT 2
- Dynamic Project Structure Tracking and Import Management [Feature Request]
- Feature request: Comparing results against other models? HOT 2
- Feature request: Run job after results to finalize
- Support Amazon Bedrock to raise rate limits HOT 2
- Update BASE_SYSTEM_PROMPT: Ensure only functions from project files are presented when checking Python files HOT 1
- Missing requirement HOT 1
- Please add support for debugging multi console programs like Textual
- Tool: edit_and_apply stops working HOT 4
- Add linux command tool
- feedback during automode
- Support for Jan/GPT4All
- Adding tags for repo HOT 1
- invalid characters when creating code
- Ollama Engineer only throws API Error HOT 1
- API Error: model "mistral-nemo" not found, try pulling it first HOT 2
- Suggestion: use `git ls-files` to ignore untracked files and possibly read entire repo into the context
- Overload issue Error code:529 HOT 1
- Claude Engineer works for very small projects, but not for big ones
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from claude-engineer.