Dynamic State Enrichment
Context shouldn't be static. Inject new files, diffs, API responses, or real-time logs directly into your compressed State Object without re-processing the entire history. Fast, cheap, and strictly up-to-date.
The Re-processing Penalty
In traditional RAG or Summarization pipelines, updating the AI's knowledge is incredibly inefficient.
The Old Way: Full Re-processing
Re-run the entire pipeline on the full history plus new data
Wait minutes (or hours) for re-compression
Burn tokens re-processing unchanged context
Pay full price for every small update
Ctxfy: Fast Forward Delta
POST only the new content (diff, schema, logs) to Fast Forward
Get an updated state in seconds, not minutes
Existing state is preserved; new context is intelligently merged
Pay only for the delta, not the full re-processing
How to Mutate State
Start with a State Object
You have a compressed state (from a previous session or API call). It represents your conversation, decisions, and constraints.
Prepare Your Delta
New data to inject: a file diff, API response, error log, schema update. Anything the model needs to know.
POST to Fast Forward
Send your base_state_id and delta_content. Ctxfy merges intelligently—no full re-compression.
Use the Updated State
Feed the enriched state to your LLM. The model has the latest context. Zero re-reading, zero waste.
Production Scenarios
Fast Forward is designed to handle noisy inputs gracefully.
Schema & API Spec Updates
Your codebase schema changed. Instead of re-explaining everything, POST the new schema to Fast Forward. The LLM's context stays in sync with your DB.
Real-time Error Logs
Debug a production issue. Grab the last 100 error lines, POST to Fast Forward. The model sees the latest failures without re-processing hours of chat history.
Agent Tool Outputs (Observation Injection)
When an agent runs a web search, it gets 10 pages of messy HTML. Don't pollute the context window! Pass the HTML to Fast Forward. Ctxfy extracts the facts and merges them into the agent's core memory.
// POST /v1/sources/{sourceId}/fast-forward
// Updates target source with new data (URI, FILE, or CONTENT)
const toolResult = await executeTool("fetch_stripe_docs");
const response = await fetch(
"https://api.ctxfy.com/v1/sources/" + sourceId + "/fast-forward",
{
method: "POST",
headers: {
"Content-Type": "application/json",
"Authorization": "Bearer " + apiKey
},
body: JSON.stringify({
type: "CONTENT",
initial_data: {
data: toolResult
},
instructions: {
}
})
});
// Fast-forward initiated. Target source updates with new context.Related features
Resume in 30 Seconds
Your job: get back to productive work. Ctxfy turns 2 hours of re-reading into one paste. Load the minimal state you need—no full history. Works with ChatGPT, Claude, Gemini, and any AI tool.
Find Any Code or Decision
Your job: reuse what you already created. Stop losing 15 minutes scrolling. Ctxfy extracts code, schemas, and decisions—searchable when you need them.
Context Hot Load
Resume work in 30 seconds—not 2 hours. Load your snapshot, paste into any AI chat, continue where you left off.
Frequently asked questions
Fast Forward allows you to update a compressed state without re-processing the entire conversation. For example, if you have a 'State Object' representing a codebase, and you change a file, you can POST the file diff to /v1/fast-forward. We intelligently merge this new context into the existing state, giving the LLM the latest information instantly.
Use Fast Forward when you have an existing state and only need to add new information (diffs, API responses, logs). Use full compression when starting fresh or when the conversation has changed significantly.
Yes. It's ideal for agents. When a tool returns large output (e.g., web search, DB query), POST it to Fast Forward instead of appending raw tokens. The agent gets dense, relevant context without blowing the context window.
Keep your context alive.
Stop fighting token limits. Start mutating state.