# What Is the Transport Budget for OpenClaw with Distilly? A Complete Guide

> Discover the OpenClaw transport budget in Distilly. Learn to optimize costs or enable intensive research by choosing between budget-friendly and budget-unfriendly settings.

- Repository: [Tianyi Zhou/distilly](https://github.com/titanwings/distilly)
- Tags: how-to-guide
- Published: 2026-09-10

---

**The transport budget for OpenClaw with Distilly defaults to "budget-friendly" for cost-efficient execution, but can be explicitly switched to "budget-unfriendly" for deeper, compute-intensive research pipelines.**

Distilly defines the transport budget through **research profiles** that control compute allocation when skills run through OpenClaw. These profiles determine whether the system executes a lightweight three-track flow or a comprehensive six-track validation pipeline, directly impacting latency and token consumption.

## Understanding the Two Transport Budget Profiles

Distilly implements a dual-profile system that governs how OpenClaw allocates computational resources during skill execution.

### Budget-Friendly Profile (Default)

The **budget-friendly** profile minimizes tool-calling steps to deliver faster, cost-effective results. This profile runs the standard three-track flow and serves as the default transport budget for all Distilly skills unless explicitly overridden. According to the source code in [`tools/skill_presets.py`](https://github.com/titanwings/distilly/blob/main/tools/skill_presets.py) (line 20), this is set via the `DEFAULT_RESEARCH_PROFILE` constant, making it the automatic selection for typical queries.

### Budget-Unfriendly Profile

The **budget-unfriendly** profile executes a rigorous four-stage pipeline: **research → audit → synthesis → validation**. This expanded workflow maps to the six-track flow described in [`SKILL.md`](https://github.com/titanwings/distilly/blob/main/SKILL.md) (lines 415-523) and consumes significantly more compute resources. This profile activates when users demand higher confidence levels or explicitly request deeper investigation capabilities.

## How the Transport Budget Is Configured in Code

The transport budget mechanism spans three critical components in the Distilly repository:

1. **Profile Definitions ([`tools/skill_presets.py`](https://github.com/titanwings/distilly/blob/main/tools/skill_presets.py))**: Line 20 establishes `budget-friendly` as the `DEFAULT_RESEARCH_PROFILE`, while the file maintains preset dictionaries defining parameters for both budget profiles.

2. **Manifest Propagation ([`tools/skill_writer.py`](https://github.com/titanwings/distilly/blob/main/tools/skill_writer.py))**: When generating skill manifests, this module inserts the selected research profile into the JSON configuration (lines 555-563), ensuring OpenClaw receives the correct budget instructions at runtime.

3. **Pipeline Execution Logic ([`SKILL.md`](https://github.com/titanwings/distilly/blob/main/SKILL.md))**: The documentation at lines 102-105 distinguishes between the budget-friendly three-track flow and the budget-unfriendly six-track flow (referencing the "celebrity" sections), detailing exactly how each profile maps to specific validation tracks.

## Setting the Transport Budget in Practice

Override the default transport budget using command-line flags or manifest configuration:

```bash

# Install with default budget-friendly profile

python3 tools/install_openclaw_skill.py --force

# Explicitly request high-compute budget-unfriendly execution

distilly run --skill celebrity --research-profile budget-unfriendly

```

For programmatic control within skill manifests:

```json
{
    "research_profile": "budget-unfriendly",
    "toolchain": {
        "quality_profile": "budget-unfriendly"
    }
}

```

The [`tools/install_openclaw_skill.py`](https://github.com/titanwings/distilly/blob/main/tools/install_openclaw_skill.py) script respects these profile settings when deploying Distilly skills into OpenClaw workspaces, ensuring the transport budget propagates correctly through the installation pipeline.

## Summary

- **Default Behavior**: OpenClaw uses the `budget-friendly` transport budget automatically, defined in [`tools/skill_presets.py`](https://github.com/titanwings/distilly/blob/main/tools/skill_presets.py) at line 20.
- **Profile Options**: Two profiles exist—`budget-friendly` (three-track, low cost) and `budget-unfriendly` (six-track with validation stages).
- **Switching Modes**: Users activate the expensive profile via `--research-profile budget-unfriendly` CLI flags or explicit manifest configuration.
- **Implementation Path**: The budget flows from [`skill_presets.py`](https://github.com/titanwings/distilly/blob/main/skill_presets.py) → [`skill_writer.py`](https://github.com/titanwings/distilly/blob/main/skill_writer.py) (lines 555-563) → OpenClaw pipeline execution as documented in [`SKILL.md`](https://github.com/titanwings/distilly/blob/main/SKILL.md).

## Frequently Asked Questions

### What is the default transport budget for OpenClaw with Distilly?

The default transport budget is **budget-friendly**, as implemented in [`tools/skill_presets.py`](https://github.com/titanwings/distilly/blob/main/tools/skill_presets.py) via the `DEFAULT_RESEARCH_PROFILE` constant at line 20. This setting automatically applies to all skill executions unless the user explicitly requests otherwise.

### How do I switch to the budget-unfriendly transport budget?

Pass the `--research-profile budget-unfriendly` flag when running a skill via CLI, or set `"research_profile": "budget-unfriendly"` in the skill's manifest JSON. This triggers the six-track validation pipeline instead of the standard three-track flow.

### Where is the transport budget logic defined in the Distilly source code?

The primary definitions reside in three files: [`tools/skill_presets.py`](https://github.com/titanwings/distilly/blob/main/tools/skill_presets.py) (default constants), [`tools/skill_writer.py`](https://github.com/titanwings/distilly/blob/main/tools/skill_writer.py) (manifest injection at lines 555-563), and [`SKILL.md`](https://github.com/titanwings/distilly/blob/main/SKILL.md) (pipeline specification at lines 102-105 and 415-523).

### What is the difference between the three-track and six-track flows?

The **three-track flow** (budget-friendly) executes minimal validation steps for speed and cost efficiency. The **six-track flow** (budget-unfriendly) runs comprehensive stages including research, audit, synthesis, and validation, providing higher confidence results at significantly higher computational cost.