Leadpoet

<h1 align="center">Leadpoet</h1>

<p align="center">
  <strong>AI sales intelligence, built on Bittensor.</strong>
</p>

<p align="center">
  <a href="https://discord.gg/tMcmbPKvz"><img alt="Discord" src="https://img.shields.io/badge/Discord-Join-5865F2?style=flat-square"></a>
  <a href="https://subnet71.com"><img alt="Dashboard" src="https://img.shields.io/badge/Leaderboard-subnet71.com-e8c76d?style=flat-square"></a>
  <a href="https://leadpoet.com"><img alt="Website" src="https://img.shields.io/badge/Website-leadpoet.com-f3f4f6?style=flat-square"></a>
  <a href="https://x.com/subnet71"><img alt="Subnet X" src="https://img.shields.io/badge/X-@subnet71-000000?style=flat-square"></a>
  <a href="https://x.com/LeadpoetAI"><img alt="Leadpoet X" src="https://img.shields.io/badge/X-@LeadpoetAI-000000?style=flat-square"></a>
</p>


---

# Leadpoet Open Source Agent Competition

Leadpoet is Bittensor subnet 71. Miners improve an open sales-research agent. The public baseline lives in the promoted `lab` branch of [pydantic-harness](https://github.com/leadpoet/pydantic-harness).

## Daily schedule

Day 0: submit models while that day's 20 ICPs stay hidden. At about **00:00 UTC on Day 1**, submissions close and evaluation starts for the baseline and prior-day models. All 20 are processed without fixed pauses between batches. Code and aggregate final scores publish when evaluation finishes. For rounds frozen with `after_scoring_day2_v1`, the benchmark stays private until the round is terminal and the submission cutoff is at least 24 hours old. Published outputs, run results, and per-ICP scores use the same boundary. Cancelled-round results and source remain private. Earlier rounds without this marker keep their original timing.

## Install and register

Use Python 3.11. Install the miner runtime:

```bash
git clone https://github.com/leadpoet/leadpoet.git
cd leadpoet
python3.11 -m venv venv
source venv/bin/activate
python -m pip install -r requirements.txt
python -m pip install -e .
```

The runtime pins Bittensor 10.5.0. Use a separate environment for the current v11 wallet CLI so its dependencies do not change the miner runtime:

```bash
deactivate
python3.11 -m venv .venv-bittensor11-cli
source .venv-bittensor11-cli/bin/activate
python -m pip install "bittensor==11.1.0"
btcli wallet create -w miner -H default
btcli wallet registrations -w miner --network finney --netuid 71
btcli tx burned-register --netuid 71 --hotkey miner/default -w miner --network finney --dry-run
```

Fund the coldkey, review the registration cost, then run the last command without `--dry-run` to register. Keep the seed phrase private. See the official [wallet guide](https://www.bittensor.com/docs/concepts/wallets) and [registration guide](https://www.bittensor.com/docs/tx/burned-register).

Return to the miner environment:

```bash
deactivate
source venv/bin/activate
```

## Build your agent

Fork the public baseline, or clone its promoted version as a starting point:

```bash
git clone --branch lab https://github.com/leadpoet/pydantic-harness.git ../my-agent
```

In `harness.py`, define or re-export this synchronous function with exactly one positional parameter:

```python
def run_icp(icp: dict) -> list[dict]:
    """Return at most five companies."""
```

Keep `LICENSE` beside `harness.py`. It must contain the complete
[Tyche AGPL-3.0 license](https://github.com/gzaentz/tyche/blob/main/LICENSE).
The CLI checks the full license text before upload, and the gateway checks the
uploaded archive again before it accepts the submission.

Change the harness, model, prompts, and approved API routing. List Python dependencies in `requirements.txt`: package names and version constraints only, with binary wheels available. URLs, VCS dependencies, local paths, nested requirements, and source builds are not supported. Return at most five companies as a JSON list. Use `[]` if there are no valid matches.

Each ICP execution attempt has a **five-minute wall-clock limit** and quotas of **60 OpenRouter, 30 Deepline, and 30 Scrapingdog calls**. The sandbox blocks direct network access. Keep the baseline's broker transport when changing the harness, or implement the same [broker protocol](lab_arena/shim.py) using the [approved operations](lab_arena/operations.py).

The output below shows the exact supported fields. It is a format example, not a real company claim. `company_linkedin`, `company_stage`, and `state` may be empty; `required_attribute` may be `null` when it is not required. `matched_icp_signal` is the zero-based position in the input's `intent_signals` list.

<details>
<summary>Output</summary>

```json
[
  {
    "company_name": "Example",
    "company_website": "https://example.com/",
    "company_linkedin": "",
    "industry": "Software",
    "employee_count": "51-200",
    "company_stage": "Series A",
    "country": "United States",
    "state": "",
    "intent_details": "Example launched its workflow platform on August 20, 2026. The launch could create more implementation work as customers adopt the platform. This activity makes Example relevant to the ICP for business software companies expanding their workflow offering.",
    "intent_signals": [
      {
        "matched_icp_signal": 0,
        "description": "Example launched its workflow platform.",
        "date": "2026-08-20",
        "url": "https://example.com/news"
      }
    ],
    "required_attribute": {
      "text": "Required characteristic.",
      "passed": true,
      "evidence_url": "https://example.com/about",
      "evidence_quote": "Supporting quote.",
      "explanation": "Why it passes."
    }
  }
]
```
</details>

Rounds that announce `intent_details_policy: "intent_details_v1"` use output
schema v5. They require one plain company-level `intent_details` paragraph and
do not accept the older `fit_summary`, `fit_evidence_urls`, signal `why_now`, or
signal `snippet` fields. The scorer verifies the paragraph against the
independently verified signals before the