Controlling cost

Where Aimogen Pro spends money, how to see it per request, and every setting that caps spending across content generation, chatbots and media.

Aimogen Pro can spend a lot of money quickly if nothing constrains it. A bulk rule left running with an expensive model and a generous token budget is the classic way to be surprised by an invoice.

Where the money goes#

Activity	Cost driver
Article generation	Input plus output tokens, multiplied by the number of stages per post
Chatbot	Tokens per message, plus the injected context on every turn
Embeddings	Tokens indexed, once per document, plus a small query cost per retrieval
Image generation	Per image, and much more per image than a paragraph of text
Video generation	Per video. By far the most expensive per unit
Text-to-speech	Per character. Cheap per request, expensive per article
Transcription	Per minute of audio
Agents	Every step is a request. A twenty-step agent is twenty requests
Internet access	A search plus several page fetches, then a larger prompt

The one people underestimate is stages per post. A bulk rule that generates a title, an outline, six sections, an introduction, a conclusion, an excerpt, categories, tags, an SEO title and an SEO description makes fourteen requests for one post — before images.

See what you actually spend#

System & Logs › Usage Logs records one row per request: user, model, mode, units, estimated price, environment and time. Usage Graphs plots the same data.

Enable it under Settings › Statistics:

Setting	Effect
`Enable Usage Tracking`	Records the row. Leave this on
`Enable Detailed Usage Tracking`	Also stores prompt and response text in `aiomatic_logmeta`

Prices shown are the plugin estimate from its own model price table. Treat the provider dashboard as authoritative.

Capping spend#

Provider-side caps#

Set a hard monthly limit in each provider dashboard. This is the only control that cannot be bypassed by a plugin misconfiguration. Do this first.

Plugin usage limits#

System & Logs › AI Usage Limits and Text-to-Speech Limits cap consumption per user and per guest.

Setting	What it does
`Enable Global Usage Limits`	Master switch
`Max User Credits` plus type	A number, counted as queries, tokens or price, for logged-in users
`Time Frame`	The period the allowance covers
`Absolute Timeframe`	Absolute means "today"; relative means "the last 24 hours"
`Max Guest Credits`	The same for visitors who are not logged in
`Full Access Users`	Users exempt from limits
`Message When Limit Reached`	Shown to the user, separately for logged-in and guest

Per-role and per-membership-plan rules are also available. See Usage limits and credits.

Model choice#

The largest single lever. Flagship models can cost ten to fifty times a mid-tier model per token.

A practical split:

Bulk articles: a mid-tier model
Titles, excerpts, meta descriptions: the cheapest capable model — these are a few dozen tokens
Flagship content: the expensive model, deliberately, on the pieces that matter
Chatbot: a fast mid-tier model. Latency matters more than eloquence

See Selecting models.

Token budgets#

Lower Max Tokens where long output is not needed. Titles do not need 2048 tokens.

Lower the context limits under Advanced AI Settings to force prompt truncation before an enormous prompt is billed. See Tokens, context and output length.

Rule volume#

Every rule has a posts-per-run setting and a schedule. Multiply them honestly:

posts_per_run × runs_per_day × requests_per_post × avg_tokens × price_per_token

Start every new rule at one post per run and only raise it once you have seen the cost per post in the usage logs.

Images#

Images cost far more than text per unit.

Generate one featured image per post, not one per section, unless you need them
Consider Cloudflare Workers for no per-image cost
Consider royalty-free sources, which are free
Raise Delay Between API Requests For AI Image Generator APIs (ms) to avoid paying for retried requests

Internet access and retrieval#

Both inflate the prompt on every call. Enable them per context rather than globally, and reduce Number Of Search Results To Add and the retrieval top-K if you do not need many.

The Batch API#

For work that can wait, the OpenAI Batch API is roughly half price. Suitable for generating a large backlog; not for anything interactive.

Agents#

Cap Max Steps per agent. An agent with a 100-step budget and a vague task will use the budget.

A cost-conscious baseline#

Settings > Statistics
  Enable Usage Tracking:                     on
  Enable Detailed Usage Tracking:            off in production

Settings > AI Writer
  Max Tokens:                                2048

System & Logs > AI Usage Limits
  Enable Global Usage Limits:                on
  Max Guest Credits:                         a low number
  Max User Credits:                          sized to your plan

Each rule
  Posts per run:                             1 to start
  Model:                                     mid-tier
  Images:                                    one featured image

Provider dashboards
  Monthly spend cap:                         set on every provider

Still stuck? Open a support ticket and include the diagnostics from Aimogen Pro › System & Logs › System Info.