Updated public guide -

Master the AIChemy workflow.

Learn the execution-oriented improve pipeline, provider capability model, workload budgets, web and extension contract, and the controls that turn rough intent into production-ready prompts.

Intent

Start with the real job, constraints, audience, and risk so the improved prompt preserves purpose instead of only sounding cleaner.

Mastery console

A living map of the improve engine

Signal
01
Intent

AIChemy starts by reading the job behind the prompt so the result preserves what you actually need done.

Hover or tap any signal to inspect how the guide translates into real product behavior.

Overview

What AIChemy does now

AIChemy improves prompts for downstream AI systems with an execution-oriented pipeline. Instead of only polishing wording, it classifies request intent, budgets context and reasoning work, applies provider-aware rendering, and returns a normalized result shape that both the web app and extension understand.

Intent first

Intent-aware improvement

AIChemy classifies whether your input is a task, a prompt draft, or a meta-prompt about improving another prompt. Explicit controls override inference when you need precision.

Bounded runs

Workload-aware execution

The active improve flow derives a workload tier from the task, risk, and requested effort so provider output, reasoning, timeout, retry, and context budgets stay bounded.

One contract

Shared result contract

Web and extension consume the same normalized result shape: improvedPrompt, techniques, explanation, warnings, workload metadata, execution score, and legacy score compatibility.

Quick start

Get productive quickly

  1. 1

    Sign in with Google and open the dashboard workspace.

  2. 2

    Connect at least one provider in Settings. Web provider credentials are stored locally in encrypted browser storage and sent only when you submit an improve request.

  3. 3

    Paste the prompt you want to improve, pick a model, choose an agent, and optionally set segment, prompt effort, or structured context assets.

  4. 4

    Use advanced controls when you want to force intent: input mode, downstream target, provider preference, desired output, target audience, project context, or risk level.

  5. 5

    Run the improvement, inspect the improved prompt, execution score, metadata, and warnings, then save it to history if you want to reuse it later.

Mastery map

Think like a power user

Interactive runbook

Prompt mastery relay

Move through the core operating loop and watch the active phase update like a command center instead of reading a static checklist.

Active phase
01 / 04
Reactive
Intent design

Frame the job

Write the real task, not a polished slogan. AIChemy works best when it can see the desired outcome, audience, constraints, and risk.

Frame the job

Write the real task, not a polished slogan. AIChemy works best when it can see the desired outcome, audience, constraints, and risk.

Choose the right controls

Use agent, model, segment, and prompt effort for broad steering. Add advanced controls only when the output needs non-negotiable intent.

Inspect the result

Read warnings, execution score, techniques, workload metadata, and explanation. Treat them as a map of what the system preserved, inferred, capped, or downgraded.

Reuse what works

Save strong outputs to history, start from templates for repeated work, and refine with project context instead of rebuilding from scratch.

Workspace

Controls that matter

Always-visible controls

Prompt inputAgent selectionModel selectionSegment selectionPrompt effortContext assets

Advanced improve controls

Input modeDownstream targetProvider preferenceDesired outputTarget audienceProject contextRisk level
Use advanced controls only when you want to force intent. If you leave them empty, AIChemy infers safe defaults from your input and surfaces warnings when confidence, capability, context, or workload limits affect the result.

Providers

Current provider model

OpenAI

Strong general, coding, and reasoning coverage. Current improve flow uses provider-specific Responses handling, normalized terminal events, and bounded recovery when a response ends because it hit an output limit.

Anthropic

Reasoning and agentic workflows with provider-specific thinking support. Improve rendering uses Anthropic-style structured sections and caps thinking budgets through the same workload policy used by every route.

Google

Gemini improve requests use API-key authentication for the current generate-content path. Provider-specific rendering and thinking options are applied without treating OAuth-only state as usable Gemini improve credentials.

Mistral / Groq / Meta

Supported through chat-completions compatible adapters and the same normalized capability matrix. Partial or length-limited completions are rejected or retried instead of being accepted as successful prompts.

The live model catalog is served through /api/models. Specific model names, tiers, streaming support, and thinking support can change over time, so treat the settings panel and public model API as the source of truth.

Extension

How the extension fits in

Popup flow

  1. 1Select text or use the popup entry path to create a pending prompt.
  2. 2If you are authenticated, the popup can improve immediately through /api/improve/stream.
  3. 3Streaming responses expose progress, provider thinking when available, and final normalized results; raw provider marker text is withheld until the final parsed result.
  4. 4Improved prompts and related metadata can be saved into extension history and synced into the web dashboard view.

Storage and safety

The extension stores sensitive local state through Web Crypto AES-GCM helpers. It validates sender identity for internal messages, restricts external messaging to official AIChemy origins, and uses server-signed extension sessions instead of unsigned local tokens.

History & templates

Saved work and reuse

History

Web and extension history preserve the normalized improve result shape, including warnings, execution score, input mode, downstream target, context metadata, workload metadata, and provider metadata.

Templates

Templates help you start faster, then advanced controls add stronger constraints when the task needs richer intent and output requirements.

Security

Important behavior and limits

  • Your web-side AI provider credentials are stored locally in encrypted browser storage and are sent only with the improve request you submit.
  • The extension stores sensitive settings locally with encryption helpers rather than plain local storage values.
  • Streaming improve responses use comment heartbeats and final-only prompt delivery: provider text is parsed on the server before the final result is sent.
  • When effort, context, thinking, output, or provider capability is capped, AIChemy returns warnings instead of pretending full parity exists.

Troubleshooting

Common issues

No providers available

Open Settings and connect a provider first. The workspace and extension both rely on a configured model before an improve request can run.

The model is available in one place but not another

The public model list comes from /api/models. If an admin changes model availability, make sure the public catalog and admin source stay aligned.

The improved prompt is too generic

Set explicit advanced controls: choose the input mode, set a more specific downstream target, and add desired output / project context so the renderer has real constraints to preserve.

A request times out or downgrades capability

AIChemy emits structured warnings when effort, context, output, thinking, or model capability is capped. Switch models, reduce context, or lower requested effort if needed.

Need more help? Open the workspace, review your current model/provider settings, and then contact support if the issue persists.