4 New MCP Servers Closing East Africa's Coordination Gap

Four new open-source MCP servers shipped this week, each targeting a different coordination failure in East Africa. Together they extend the stack to cover credit, markets, insurance, and reputation — the core systems that underpin economic participation in mature economies.

What Was Missing

The gap isn't primarily one of wealth. It's one of coordination infrastructure: the systems that let strangers transact, workers prove their skills, farmers price their crops, and individuals access capital.

Each of these has a technology-compressible solution.

mkopo-mcp — Alternative Credit Scoring

The problem: FICO scores took 50 years of credit card history to build. Most Kenyans have no formal credit history at all.

The insight: M-PESA is a de facto financial history. Regular income deposits, utility payment regularity, savings behaviour, Fuliza usage patterns — all predict repayment with measurable accuracy.

pip install mkopo-mcp    # coming soon

Five tools:

• alternative_credit_score — 300–850 score from M-PESA behavioural signals
• mpesa_creditworthiness — cash flow analysis and loan capacity
• credit_report_summary — structured report like a loan officer reviews
• loan_eligibility — match to appropriate Kenya lender types
• credit_improvement_tips — personalised steps by credit tier

→ github.com/gabrielmahia/mkopo-mcp

soko-mcp — Commodity Price Intelligence

The problem: A maize farmer in Nakuru doesn't know that Nairobi prices are 40% higher that week. Traders know. Farmers don't. Information asymmetry is a direct tax on rural producers.

pip install soko-mcp    # coming soon

Five tools including sell_hold_decision:

sell_hold_decision(
    commodity="maize",
    market="nakuru",
    quantity_bags=50,
    storage_cost_per_month_kes=500,
    months_can_store=3
)
# → HOLD 2 MONTHS — expected KES 8,500 net gain after storage

→ github.com/gabrielmahia/soko-mcp

bima-mcp — Insurance Intelligence

The problem: Kenya insurance penetration is 2.3% of GDP. The barrier is distribution, claims verification, and actuarial data — all technology-compressible.

pip install bima-mcp    # coming soon

Six tools covering NHIF coverage queries, parametric crop risk scoring (NDMA + ACRE Africa methodology), microinsurance comparison, and chama pool sizing.

→ github.com/gabrielmahia/bima-mcp

sifa-mcp — Portable Reputation & Skills Passport

The problem: A skilled mason in Mombasa cannot prove her track record to a contractor in Nairobi. Economic mobility requires reputation portability. Without it, hiring defaults to personal networks.

pip install sifa-mcp    # coming soon

Six tools:

• create_worker_profile — portable profile with skills and location
• add_service_record — log completed jobs with client ratings
• get_reputation_profile — full trust score for employers, SACCOs, banks
• generate_reference — formatted professional reference letter
• compare_workers — rank candidates by trust score and skill match

# A mason's profile follows her to any market
create_worker_profile(
    full_name="Amina Odhiambo",
    primary_skill="construction",
    sub_skills="masonry,tiling,plastering",
    location="Mombasa",
    years_experience=8
)
# Trust tier: ESTABLISHED after 5+ verified records

→ github.com/gabrielmahia/sifa-mcp

The Coordination Stack So Far

All MIT licensed. All DEMO data clearly labelled. All deployable against any MCP-compatible AI client.

Full portfolio: gabrielmahia.github.io

All demo data for educational purposes. Nothing in this article constitutes financial, legal, or insurance advice.