Prompt Template Library
Keep reusable prompts in one place instead of rewriting the same instructions each time.
What it adds
A managed collection of named prompt templates, scoped by owner and role, that the app's AI features run from.
What your agent is told to do
5
What your agent is told to do
5-
1
Find every place in the app where prompt text is currently assembled inline, and move that text into named templates that the feature loads at run time. Leaving one copy behind guarantees the two drift apart.
-
2
Split each template into a system portion, which only administrators can edit and which users never see, and a task portion that authors may change. The system portion is where safety rules, output shape, and tone live, and a user editing it can disable them by accident.
-
3
Scope templates to a user, a workspace, or a role using the app's existing permission model, and default a new template to its author rather than to everyone.
-
4
Record who created and last changed each template, when, and which features use it, so an unused template can be retired and a broken one can be traced back to a change.
-
5
Placeholder syntax, typing, and validation of substituted values are owned by Prompt Variables; this feature owns storage, scoping, and discovery. Do not build a second substitution mechanism here.
Edge cases it handles
8
Edge cases it handles
8- System instructions and user-editable task content must be stored and rendered separately. A single editable blob lets an author delete the safety and formatting rules without realising the feature depended on them.
- Templates must respect the app's scoping rules, so a workspace template never appears in another workspace and a role-restricted template never appears to users who cannot run it.
- A template with unfilled required variables must be blocked before any request is sent, because a prompt containing literal placeholder text produces confidently wrong output rather than an error.
- A template that references data or a tool the running user cannot access must fail with a clear explanation. Silently dropping the inaccessible part produces an answer built on a gap nobody was told about.
- Ownership and last-updated information must be recorded and displayed, or a library of a hundred templates becomes impossible to audit or prune.
- Deleting or renaming a template that a live feature depends on must be refused or must warn with the list of dependents, rather than breaking the feature at the next run.
- Templates must have a length ceiling, because an unbounded template multiplied by every request is a cost problem before it is a quality problem.
- If the template store is unreachable, the feature must fail with a clear message rather than falling back to an empty prompt.
Definition of done
8
Definition of done
8- All prompt text used by the app lives in the template library, with no inline copies remaining.
- System instructions are separated from user-editable content and are not visible or editable to ordinary authors.
- Templates are scoped by user, workspace, and role, and new templates default to their author.
- Each template shows its owner, last editor, last update, and the features that use it.
- Deleting or renaming a template in use warns with its dependents rather than silently breaking them.
- Running a template with missing required values is blocked before any request is made.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
Multi-Model Routing
Multi-Model Routing
Send each AI request to the right model using rules you can read and test.
What it does
A deterministic routing layer that picks a model per request from task type, context size, latency budget, and data sensitivity.
How it works
- 1 Express routing as explicit, ordered rules over inputs the app can measure: task type, estimated context size, latency budget, and the sensitivity classification of the data involved. A rule set that can be read line by line can be reviewed and tested.
- 2 Make routing deterministic. The same inputs must always produce the same route, so a bad output can be reproduced and a rule change can be evaluated. Randomised or load-based selection turns every incident into guesswork.
- 3 Classify data before routing and refuse to route restricted content to any destination not approved for it. This check is a hard block, not a preference, and it must run before the request is assembled.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/multi-model-routing
AI Cost Budgets
AI Cost Budgets
Cap what AI features are allowed to spend before the bill arrives.
What it does
Monetary spending limits on AI work, scoped by workspace, feature, and time period, enforced before a run starts.
How it works
- 1 Find every place the app calls a model and route all of them through one accounting point that records estimated and actual spend against a scope. A budget that only covers the chat feature is not a budget.
- 2 Estimate the cost of a run from the size of its input before dispatching it, and refuse anything that would exceed the remaining budget on its own.
- 3 Reserve the estimate against the budget when the run starts, then reconcile to the real usage figures when it finishes, releasing whatever was over-reserved.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/ai-cost-budgets
Model Selection
Model Selection
Let each AI task run on the model that suits its quality, speed, and cost needs.
What it does
A per-task model choice, drawn from the models the app already has configured, with capability filtering and safe defaults.
How it works
- 1 Enumerate the models the app already has access to and record what each one can actually do: context capacity, whether it can return the structured output the task requires, whether it supports the tools the task calls, and its relative cost and speed.
- 2 Offer only the models that satisfy the task's requirements. A task that needs structured output must not list a model that cannot reliably produce it, because the failure appears later as malformed responses rather than as an unavailable option.
- 3 Store the choice against the specific task, not as one global setting. A single default forces a summarisation task and a classification task onto the same tier when they have opposite needs.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/model-selection
How it works
-
1
Copy the link
Grab the Markdown instruction URL for this feature.
-
2
Give it to your AI
Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
-
3
It inspects, then implements
Your agent reads your existing app first, then adds the feature to fit it.
Works with your stack
These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.
Need it tighter than that? Customize the feature and tell it exactly what you're running.