Blockchain Token Performance Optimization

$10-30 CAD

Completed

Posted 6 days ago

$10-30 CAD

Paid on delivery

I would like to track the usage of tokens and their cost at user level . In our office We have aws bedrock llms , specifically we have got access to anthropic claude through aws bedrock . There are multiple ways people at office are accessing these llms . some through CLI, some using APi and some through agents . I need a straightforward way to count the token usage and cost at user level , application level and agent level. The objective is to track and restrict the users at our office . The metrics can then be pushed to datadog for observability.

Application Performance Monitoring

AI Model Integration

Project ID: 40482049

About the project

2 proposals

Remote project

Active 6 days ago

Looking to make some money?

Email address

Benefits of bidding on Freelancer

Set your budget and timeframe

Get paid for your work

Outline your proposal

It's free to sign up and bid on jobs

Awarded to:

@anirudhat

Hi, I've built similar token tracking and cost attribution systems on AWS Bedrock, so I understand exactly what you need — a unified metering layer across CLI, API, and agent access points that captures per-user, per-application, and per-agent token consumption with associated cost calculations. My approach would be to implement a lightweight proxy/middleware layer that intercepts all Bedrock API calls, extracts token usage from response metadata, maps it to authenticated users via IAM or API key identity, calculates costs based on Claude model pricing tiers, and enforces configurable usage quotas for restriction. I'll set up a structured pipeline to push these metrics into Datadog with pre-built dashboards for real-time observability. I can start immediately and have done similar integrations with Datadog custom metrics before.

$10 CAD in 1 day

5.0

(1 review)

3.9

@MahiWorksGreat

I can handle this for $150 CAD. I have experience with Python, AWS-based workflows, APIs, observability, and automation systems. For your Bedrock setup, I would implement a centralized tracking layer that captures token usage, model calls, and associated costs across CLI users, API consumers, and agent-based workloads. The solution would attribute usage at user, application, and agent level while providing clear reporting and budget controls. The approach would leverage Bedrock invocation logs, CloudWatch/CloudTrail telemetry, custom middleware where required, and cost-calculation logic based on Claude pricing. Metrics can then be exported to Datadog for dashboards, alerts, quotas, and usage restrictions. The end result would be a reliable usage-monitoring framework that gives complete visibility into LLM consumption across your organization.

$150 CAD in 3 days