> For the complete documentation index, see [llms.txt](https://docs.bardiel.tech/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.bardiel.tech/roadmap.md).

# Roadmap

## Roadmap

### Comprehensive Roadmap

Bardiel’s mission is to become a **trust and execution layer for Base, Virtual, ERC-8004, and other onchain agent ecosystems** — the service layer agents, apps, and markets call when they need reliable compute, validation, factchecking, or dispute resolution instead of just another answer.

Bardiel is:

* **Built by Virtual** – with deep native integration into GAME, ACP, and the broader Virtual agent ecosystem.
* **Powered by Cortensor** – using Cortensor’s decentralized inference and PoI/PoUW validation rails as its execution and proof backend.

The long-term vision is:

* Any **Base-native agent or app** can call Bardiel when it needs trusted execution or a safety checkpoint before acting.
* Any **Virtual agent or ACP market** can call Bardiel when it needs validation, factchecking, or optional dispute support.
* Any **ERC-8004-compatible registry, market, or agent flow** can use Bardiel as a validator / oracle surface.
* Bardiel itself becomes **data-informed** over time, learning better policies and scoring models from real agent interactions.

As with Cortensor, this journey is complex and experimental. We expect to iterate based on:

* real agent traffic and validation data
* evolving Cortensor capabilities
* feedback from builders across Base, Virtual, ERC-8004, and other agent ecosystems

We ask for your patience and input as we refine Bardiel’s behavior, economics, and integration surfaces.

***

### Technical Roadmap (Updated 2026 Focus)

For now, Bardiel’s public roadmap is primarily focused on **2026**, with milestones shifted roughly **one quarter later** than the original plan and the **token launch targeting Q2 2026**.

Core milestones:

* **v0.1 (target: Q2 2026)**\
  Spec, token genesis, and a stable testnet trust agent running on Cortensor.
* **v0.2 (target: Q3 2026)**\
  Data-driven validation engine with tiered policies (`fast`, `safe`, `oracle`, `adaptive`) and early scoring models.
* **v0.5 (target: Q4 2026)**\
  Bardiel used as a default trust and execution layer across major supported flows, especially Virtual GAME / ACP and early broader ecosystem integrations.
* **v1.0 (target: early 2027)**\
  Full arbitration flows and multi-step / trajectory reasoning checks, suitable for higher-stakes Base-native, ACP, ERC-8004-compatible, and other onchain agent workflows.

You can find the detailed breakdown here:

**👉** [**Technical Roadmap: 2026 – Bardiel as a Trust & Execution Layer**](https://docs.bardiel.tech/roadmap/technical-roadmap-2026-bardiel-as-virtuals-trust-agent)

***

### Phase A: Prototype & Data Foundation (Now → Q2 2026)

**Objective:** Make Bardiel genuinely useful on **Cortensor Testnet** and start collecting data on real agent interactions.

**Focus Areas:**

* Implement Bardiel as a callable trust and execution surface for early integrations.
* Early experiments across:
  * **Virtual GAME / ACP**
  * **Base-native agent workflows**
  * **x402 payment flows**
* Logging of tasks, miner outputs, and Bardiel verdicts.
* First validation and scoring experiments on real traffic.

> Note: This phase was originally scoped around Q1 2026 and is now effectively pushed into **Q2 2026**.

***

### Phase B: Bardiel 0.x – Trust Engine for Supported Ecosystems (Q2 → Q4 2026)

**Objective:** Turn Bardiel from a thin wrapper over Cortensor into a proper **trust engine** with policies, scoring, and ecosystem-ready integrations.

**Focus Areas (with updated timing):**

* **v0.1 (Q2 2026)**\
  Spec, **token launch in Q2**, and a stable testnet trust agent.
* **v0.2 (Q3 2026)**\
  Tiered validation (`fast`, `safe`, `oracle`, `adaptive`), scoring models, and early policy tuning based on Phase A data.
* **v0.5 (Q4 2026)**\
  Bardiel used as a default trust layer for many serious workflows, especially:
  * Virtual GAME / ACP
  * early Base-native agent apps
  * early ERC-8004-compatible validator / oracle patterns

The intent is that by the end of Phase B, most “serious” integrations can route through Bardiel for delegation and validation by default, instead of relying on ad hoc execution or trust checks.

***

### Phase C: Bardiel 1.x – Arbitration & Long-Horizon Trust (Early 2027 → Beyond)

**Objective:** Extend Bardiel beyond single-task checks into **full dispute arbitration** and **trajectory-level validation** for long-running agents.

**Focus Areas:**

* **v1.0 (target: early 2027)**\
  Oracle-grade arbitration for selected markets and higher-stakes integrations across Base-native, Virtual / ACP, and ERC-8004-compatible flows.
* Chain-of-thought / trajectory audits for complex agents (multi-step, long-horizon jobs).
* Using Bardiel verdicts as **long-lived trust artifacts** across supported ecosystems, including:
  * reusable evidence for agent markets
  * receipts / proof objects for settlement and audit
  * cross-workflow trust signals over time

This phase is explicitly **pushed out by roughly one quarter** compared to the original plan, reflecting the updated v0.x schedule and Q2 token launch.

***

### Summary

Bardiel’s roadmap is intentionally:

* **Agent-first** – focused on how agents, apps, and markets experience trust.
* **Data-driven** – every phase depends on the data we gather from real usage.
* **Infrastructure-aware** – aligned with, but not dominated by, Cortensor’s own testnet / mainnet timeline.
* **Ecosystem-broad** – designed for Base, Virtual, ERC-8004, and other onchain agent ecosystems, not just a single app or market.

As Cortensor, Virtual, Base-native agent ecosystems, and standards-compatible agent markets evolve, this roadmap will be updated — but the core goal remains the same:

> **Bardiel should be the easiest, most reliable way for any agent, app, or market to ask:**\
> \&#xNAN;**“Can I trust this — and who should run it for me?”**