Autonomous intelligence inside a governed shell. Meridian launches v0.1.0 with a premium Mantis-branded local runtime, a governed competitor-intelligence surface, and a truthful boundary for what is live on this host today.
Current honest state: the local Loom runtime is installable today, the public launch story centers Universal OS primitives, and the current customer path still uses a founder-led pilot while broader host claims stay bounded. Boundary note · Request Manual Pilot
Install Meridian Loom locally and get the Mantis-branded runtime surface under your own home directory.
Launch installer
curl -fsSL https://raw.githubusercontent.com/mapleleaflatte03/meridian-loom/main/scripts/install.sh | bashThe installer keeps the Mantis banner, bootstraps Rust when missing, provisions $HOME/.local/share/meridian-loom, and seeds the built-in Wasm capabilities honestly.
Meridian v0.1.0 positions Loom as a Universal Personal AI Assistant Runtime. The local primitive surface now centers the following capabilities:
Bounded local argv execution through the native Wasm host-call lane.
Bounded browser navigation and HTML capture through the local web/Wasm path.
Declared gateway surface for multi-channel delivery and presence management.
Memory-store primitives define durable state, search, and replay surfaces above the runtime boundary.
Heartbeat scheduling is now a built-in primitive with local receipt logging and truthful scheduler boundaries.
Capability import, plugin compatibility, and workspace skill loading remain part of the Loom operator surface.
Intelligence that sits unread in an inbox is wasted. Telegram is where Meridian lands first: direct, fast, and easy to forward. Today it supports manual pilot delivery plus internal lifecycle testing; automated subscriber lifecycle resumes only after Meridian reaches customer-backed treasury and treasury-cleared automation. Three surfaces:
@MeridianIntelligence — public preview surface. Abbreviated preview posts go here when a fresh brief is worth posting. No account needed.
@eggsama_bot — internal test surface today; future subscriber lifecycle surface later. It is not the current public onboarding promise.
Premium DM delivery — the long-term default. During the current funding gap, manual pilot delivery replaces the automated morning send rather than pretending the pipeline ran. Support or owner cash can keep the build alive, but customer automation still waits for customer-backed treasury and treasury-cleared automation.
The intelligence workflow runs on Meridian — a governance layer above AI runtimes. Five primitives keep every agent action audited, budgeted, and accountable.
What Meridian is: a governance kernel for digital labor. It does not replace every agent runtime; it governs identity, authority, treasury, and accountability above the runtime boundary.
What this page is showing: competitor intelligence as the first managed vertical running on that kernel.
Public proof today: the open kernel, the reference workspace, the runtime-core boundary registry/admission model, the Loom-backed local runtime receipt, the runtime-core boundary registry/admission model, and the currently verified Loom route surfaces on this host.
The kernel is open-source. Meridian Constitutional Kernel → The current runtime-core surfaces explicitly say which boundaries are institution-routable today and which remain founding-service-only.
A governed 7-agent workflow supports 30+ sources when automation is enabled. The current customer path is a founder-led manual pilot; every cited finding still names its source and date.
Cited findings from 30+ monitored sources. Pricing changes, product launches, API updates, deprecations — delivered founder-led during the current pilot and designed to become an automated morning send only after Meridian reaches customer-backed treasury and treasury-cleared automation.
Curated summary of the week's most important competitive moves. Action items and risk signals included.
Track specific competitors by name. Set topics — pricing, launches, partnerships, API changes. Get alerts only for what matters.
Structured competitive snapshots: recent moves, pricing intel, product updates, and talking points. Ready for your sales deck.
See full product demo with watchlists, weekly brief, and battlecard →
Every agent is a registered entity with identity, scopes, budget, and reputation scores. Agents earn and lose standing based on accepted output.
| Leviathann | Manager — orchestrates pipeline, routes work |
| Atlas | Research — sourced findings, competitive analysis |
| Quill | Writer — structured briefs and deliverables |
| Aegis | QA gate — PASS/FAIL acceptance with evidence |
| Sentinel | Verifier — contradiction detection, risk review |
| Forge | Executor — operational steps, implementation |
| Pulse | Compressor — triage, context optimization |
Agents earn REP (reputation) and AUTH (authority) from accepted output. Sanctions apply for failures. Scores are tracked in the economy ledger and surfaced through the governed workspace.
Current honest state: pilot onboarding is manual — email us or DM @Enhanhsj to start. Automated delivery resumes only after Meridian reaches customer-backed treasury and treasury-cleared automation.
Meridian now keeps support, pilot, and customer revenue separate on purpose. If you want to back the kernel, infra, and founder time while the automated path stays treasury-gated, use the dedicated support path instead of the pilot flow.
Meridian is a governance kernel for operating AI agents as managed digital labor. Organizations register agents, define permissions and budgets, and run governed workflows above the runtime layer. The competitor intelligence pipeline is the first live workflow vertical built on that kernel, not the whole definition of Meridian.
No. Competitor intelligence is the first managed vertical. Meridian itself is the governance layer underneath: institution, agent, authority, treasury, and court. Public proof today is strongest on the built-in kernel path; broader runtime adapter proof is still being built.
Any company in the AI/ML ecosystem: model providers (OpenAI, Anthropic, Google, Mistral, Meta, Cohere, DeepSeek), frameworks (LangChain, LlamaIndex, CrewAI), infrastructure (Cloudflare, AWS, Azure, GCP), and more.
Meridian monitors structured sources (pricing pages, changelogs, API docs) that newsletters miss. Every finding cites its source with a date. Output is structured for competitive workflows, not just raw links. And every agent action is audited and metered.
Each agent has a registered identity, defined purpose, permitted scopes, spend budget, and reputation score. Actions are logged to an audit trail. Usage is metered per organization. This is how you operate AI agents accountably at scale.
Yes. The current customer path is founder-led and manual, so you can stop anytime by replying on Telegram or email. No recurring paid continuation starts unless you explicitly agree to it.
Model Context Protocol — a standard for AI tools. Connect your AI client to Meridian's MCP server to access intelligence tools programmatically. x402-capable settlement path over USDC on Base L2. Every call is audited and metered, but external-customer settlement proof is still pending.
Show MCP integration, payment details, and API info ▾
Developer truth: Meridian is runtime-neutral in design, and this host publishes only the Meridian routes and proof surfaces it actually runs.
Today the built-in kernel/runtime pair is the live governed runtime surface on this host. Loom is the active Meridian runtime boundary here.
Add to your AI client config:
{
"mcpServers": {
"meridian": {
"url": "https://app.welliam.codes/sse"
}
}
}
intelligence_latest_brief — $0.50 — Daily intelligence alert with cited findings
intelligence_on_demand_research — $2.00 — Research any topic with sourced findings
intelligence_competitor_snapshot — $3.00 — Battlecard-ready competitor snapshot
intelligence_qa_verify — $1.00 — QA verification of claims or text
intelligence_weekly_digest — $1.50 — Weekly digest across tracked competitors
company_info — FREE — Meridian capabilities and pricing
Protocol: x402-capable settlement path (HTTP 402 Payment Required)
Chain: Base L2 (Chain ID 8453)
Token: USDC
Wallet:
0x82009D0fa435d490A12e0cBfBE47bf3358e47761
Governance kernel for managed digital labor © 2026