fetch-agents 0.1.0

Python client for fetch.ai AI agent services (code context, context doctor, AI teleporter)

fetch-agents

Python client for the AgentsLab AI agent services.

Three agents, one SDK:

• Code Context Agent — slim, map, extract, mark code layers
• Context Doctor — compress, analyze, optimize LLM context windows
• AI Teleporter — teleport, search, merge chat export knowledge bases

Install

pip install fetch-agents

Quick Start

from fetch_agents import FetchAgents

# Register once (free, no email required)
info = FetchAgents.register(name="myapp")
api_key = info["api_key"]  # save this

client = FetchAgents(api_key=api_key)

# Slim a code file to only the layers relevant to your task
result = client.slim(
    url="https://raw.githubusercontent.com/you/repo/main/app.py",
    task="fix the auth handler"
)
print(result["context"])

# Compress LLM context to reduce tokens
result = client.compress(text="...long assistant output...", max_length=200)
print(result["compressed"])

# Teleport a chat export into a searchable knowledge base
with open("chatgpt_export.json") as f:
    data = f.read()
index = client.teleport(data)
hits = client.search(index["index_id"], query="how do I fix auth?")

API Key via Environment

export FETCH_API_KEY=fak_xxx

client = FetchAgents()  # picks up FETCH_API_KEY automatically

Free Tier

50 requests/day per key, no credit card needed. For higher volume, top up with crypto:

invoice = client.topup(amount_usd=5, currency="btc")
print(invoice["pay_url"])  # pay this address

Supported: BTC, LTC, BCH, DOGE, TRX, ETH, BNB, USDT, USDC, and more.

All Methods

Method	Agent	Description
register(name)	—	Self-service registration, returns API key
me()	—	Quota + balance
topup(amount_usd, currency)	—	Create crypto invoice
topup_status(invoice_id)	—	Check invoice status
currencies()	—	List supported currencies
slim(url, task)	Code Context	Relevant layers only
map(url)	Code Context	List all LAYER sections
extract(url, layer)	Code Context	Extract one layer
mark(url)	Code Context	Auto-insert LAYER markers
compress(text, max_length)	Context Doctor	Token-compress text
analyze(messages)	Context Doctor	Score messages by importance
optimize(messages)	Context Doctor	Remove noise, compress verbose turns
extract_memory(messages)	Context Doctor	Extract structured facts
teleport(data)	AI Teleporter	Upload chat export, create index
search(index_id, query, top_k)	AI Teleporter	Semantic search
merge(index_ids)	AI Teleporter	Merge indexes
formats()	AI Teleporter	List supported export formats

Links

• Website & docs: https://agentslab.duckdns.org
• Issues: https://github.com/SoftwareValete/fetch-agents/issues