Subscription management 3 min read
Managing multiple AI subscriptions without losing track of the work
Separate the login, the harness, and the project so several subscriptions remain understandable and useful.
Blog
Notes from the ledger.
Practical guides to agent budgets, multiple AI subscriptions, project attribution, and the costs that disappear between teams and tools.
Cost attribution 3 min read
Why enterprise AI spend becomes hard to see
Subscriptions, API bills, shared keys, and infrastructure each reveal only part of what a company spends on AI.
Browse the journal
25 articles · 2 featured above
Subscription management3 min read
An agent’s usage share is not its subscription quota
Why a percentage can fall while an agent keeps working, and how to read harness shares and account limits correctly.
Subscription management3 min read
Why $100 of subscription usage may not mean a $100 bill
Keep list-price activity, subscription fees, and any additional vendor charges separate in your AI cost reports.
Subscription management3 min read
Switching subscription accounts without losing the thread
Plan account handoffs around session state, tool execution, and quota evidence rather than treating logins as interchangeable.
Subscription management3 min read
When a long Claude Code session becomes expensive
A practical way to separate useful context from repeated reading, oversized tool output, and unproductive loops.
Cost attribution3 min read
How to attribute OpenRouter spend to the right project
Use the billed key and recorded ownership, while preserving account-level totals and unresolved usage.
Cost attribution3 min read
Shared API keys make project costs harder to explain
A shared credential can be convenient while removing the evidence needed for ownership and incident response.
Cost attribution3 min read
Reconciling outside AI spend without counting it twice
Compare provider bills with proxy records using matching periods, identities, and measurement bases.
Cost attribution3 min read
A cloud usage estimate is not the final invoice
Why infrastructure costs need a measurement basis before they can be compared with agent spending.
Budget controls4 min read
Budget alerts and hard caps solve different problems
Put the spending decision on the request path when a notification would arrive too late to stop the next call.
Budget controls3 min read
How to set a first budget for an AI agent
Use a bounded pilot, an explicit stop condition, and observed completed tasks to choose a useful starting cap.
Budget controls3 min read
Rate limits do not answer the budget question
Control request pace and monetary exposure separately, then make retry behavior respect both.
Budget controls3 min read
Choosing a budget window that matches the work
Run, daily, weekly, and lifetime limits imply different restart and escalation behavior.
Budget controls3 min read
Why concurrent requests need reservations
Prevent several requests from each spending the same apparent remainder of a tab.
Engineering3 min read
Reserve, commit, refund: the lifecycle behind a tab
Follow one metered call from an admission estimate to a final ledger outcome, including the uncertain cases.
Engineering3 min read
Give retries a budget of their own
Bound repeated requests by attempts, elapsed time, and the remaining task allowance.
Engineering3 min read
Streaming responses still need a final accounting outcome
Keep first-token latency separate from the evidence needed to settle a model request.
Engineering3 min read
Prompt caching changes the cost of repeated context
Measure cache reads, writes, and misses before deciding whether a long prompt is wasteful.
Engineering3 min read
What stays local with a self-hosted agent proxy
Understand the data plane, the hosted control plane, and the availability tradeoff before choosing a deployment mode.
Team operations3 min read
Test your agent kill switch before you need it
A small drill should prove where the stop takes effect, what happens to in-flight work, and how recovery is authorized.
Team operations3 min read
A spending cap does not replace tool permissions
Bound model consumption and the actions an agent can perform as separate responsibilities.
Team operations3 min read
What an agent communication map can actually prove
Distinguish observed sends, tool outcomes, and missing coverage before drawing conclusions from a graph.
Team operations3 min read
Measure cost per completed task, not just tokens
Connect usage to accepted outcomes so a cheap request does not hide an expensive workflow.
Team operations3 min read
A weekly AI spend review that produces decisions
Bring account capacity, project attribution, and unresolved costs into one short operating routine.
Written by FlockTab. References and worked examples in every guide.
Subscribe via RSS