Pi is INCREDIBLE - Building a Custom Coding Agent Live
In this live stream Cole Medin explores Pi, a deliberately minimal coding agent you build on top of, testing its many model providers and extensions, integrating it with his open-source Archon harness, and building a custom Archon-dispatch extension while running Kimmy K2.6 to show where cheaper models succeed and where they hit their limits.
Cole MedinWatchTranscript found
Quick learning frame
Read this before watching.
A local runtime lesson is about fit: model, quantization, hardware, endpoint, latency, privacy, tool integration, and task limits.
New playlist item from Cole Medin; queued for transcript-backed review, topic mapping, and a practical learning artifact.
Skill you build: The ability to mold a minimal coding agent to your own workflow through custom extensions and harness engineering, and to route cheaper open-source models to the parts of a workflow where they hold up while reserving a powerful model for where it's needed.
Watch for the shift from claim to mechanism. The learning value is the point where the transcript reveals a repeatable action, tool boundary, context move, review habit, or artifact.
Concept diagram
Where this video fits.
01Task
02Hardware
03Model/quantization
04Runtime endpoint
05Agent tool loop
06Benchmark task
07Fallback
Deep lesson
Turn this video into working knowledge.
16,380 cleaned transcript words reviewed across 4,654 timed caption segments.
Thesis
Pi is INCREDIBLE - Building a Custom Coding Agent Live teaches a practical local model/runtime move: In this live stream Cole Medin explores Pi, a deliberately minimal coding agent you build on top of, testing its many model providers and extensions, integrating it with his open-source Archon harness, and building a custom Archon-dispatch extension while running Kimmy K2.6 to show where cheaper models succeed and where they hit their limits.
The goal is not to remember the video. The goal is to extract the operating principle, tie it to timestamped evidence, test how far the claim transfers, and make something reusable.
0:14
Adapt Pi to you
“other coding agents right now. And so, the idea behind Pi is it is a minimal coding agent. It's made to be a coding agent that you build on top of instead of taking a really massive bloated...”
Pi is a minimal coding agent designed so you build on top of it and adapt it to your workflow instead of retrofitting a bloated tool; it supports many providers out of the box (OpenRouter, Kimmy, Miniax, Qwen, Codex, Copilot), though using an Anthropic Claude Pro/Max subscription with Pi is against Anthropic's terms of service and not recommended. Open Pi's provider list and pick one non-Anthropic model (via OpenRouter or a subscription) to configure, noting why you'd choose it over the default.
26:40
Extensions need pointing
“factory. Curious how pi could help. So the the reason pi could be really helpful for the dark factory is because I have a very specific workflow for the way that I have the coding agent evolve the...”
Pi ships minimal and installs only the extensions you need (sub-agents and MCP aren't native), but Pi didn't automatically understand which installed extensions it had, getting confused until told explicitly to use the questionnaire extension, suggesting you may need global rules that specify when to invoke a given extension. Install one Pi extension and test whether the agent uses it from a natural prompt, then add a global-rule line telling it exactly when to trigger that extension.
87:25
Limits of cheap models
“extension so that the output of like the the last output from the workflow needs to go directly back into the the PI session I thought that's what you already had set up, but obviously not. It's very...”
Building a custom Archon-dispatch extension with Kimmy K2.6, the model got the extension almost working but repeatedly failed to inject the workflow's final output back into the Pi session; Cole notes he'd bet a good amount he'd have avoided this with Opus 4.7, illustrating using a cheap model to start or research and implement, then a powerful model to fix the final bug. Take a small extension or script, draft it with a cheaper model, and note exactly where it breaks so you can hand just that final fix to a stronger model.
01
Task
Start with this video's job: In this live stream Cole Medin explores Pi, a deliberately minimal coding agent you build on top of, testing its many model providers and extensions, integrating it with his open-source Archon harness, and building a custom Archon-dispatch extension while running Kimmy K2.6 to show where cheaper models succeed and where they hit their limits. Treat "Task" as the outcome you are trying to make visible, not a topic label. Anchor it to 0:14, where the video says: “other coding agents right now. And so, the idea behind Pi is it is a minimal coding agent. It's made to be a coding agent that you build on top of instead of taking a really massive bloated...”
02
Hardware
Use "Hardware" to locate the part of the local model/runtime mechanism the video is demonstrating. Ask what changes in your real setup if this claim is true. Anchor it to 26:40, where the video says: “factory. Curious how pi could help. So the the reason pi could be really helpful for the dark factory is because I have a very specific workflow for the way that I have the coding agent evolve the...”
03
Model/quantization
Turn "Model/quantization" into the reusable artifact for this lesson: A local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule. This is where watching becomes something you can inspect and reuse.
04
Runtime endpoint
Use "Runtime endpoint" as the application surface. Decide whether the idea touches a browser flow, a local file, a model choice, a source document, a UI, or a review step.
05
Agent tool loop
Use "Agent tool loop" to prove the lesson. The evidence should connect back to the video title, transcript anchors, and a concrete output, not a generic best-practice claim.
06
Benchmark task
Use "Benchmark task" to carry the idea forward: save the prompt, checklist, diagram, or operating rule that would make the next agent run better.
07
Fallback
Connect "Fallback" to Pi is INCREDIBLE - Building a Custom Coding Agent Live by naming the claim, the evidence, and the artifact it should produce.
Example
Source-backed artifact packet
Convert the video into a scoped artifact request that includes the transcript claim, mechanism, acceptance criteria, and proof. The output should be a local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule..
Example
Local model/runtime proof brief
Separate what the speaker claims, what the demo actually proves, and what still needs outside verification before you adopt the local model/runtime pattern.
Example
Teach-back module
Transform the lesson into a definition, a Task -> Hardware -> Model/quantization -> Runtime endpoint -> Agent tool loop -> Benchmark task -> Fallback diagram, one misconception, one practice exercise, and a check-for-understanding question.
Do not learn it wrong
Treating the title as the lesson without checking what the transcript actually says.
using a local model like ChatGPT
ignoring latency/context limits
no benchmark task
Letting the lesson drift into local-model ideology.
Letting the lesson drift into hardware specs without workflow fit.
Letting the lesson drift into benchmarks unrelated to the actual task.
Do not count this as learned until these are true.
01
State the transcript-backed claim in your own words: In this live stream Cole Medin explores Pi, a deliberately minimal coding agent you build on top of, testing its many model providers and extensions, integrating it with his open-source Archon harness, and building a custom Archon-dispatch extension while running Kimmy K2.6 to show where cheaper models succeed and where they hit their limits.
02
Explain the practical stakes without hype: New playlist item from Cole Medin; queued for transcript-backed review, topic mapping, and a practical learning artifact.
03
Map the idea onto the Task -> Hardware -> Model/quantization -> Runtime endpoint -> Agent tool loop -> Benchmark task -> Fallback sequence and name the weakest link.
04
Produce the artifact and include the evidence that proves it: A local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule.
Put it into practice
Give this grounded prompt to Codex or Claude after watching.
You are helping me turn one specific YouTube video into real, durable learning.
Source video:
- Title: Pi is INCREDIBLE - Building a Custom Coding Agent Live
- URL: https://www.youtube.com/watch?v=lK9o5Wu2upU
- Topic: Agent Architecture
- My current learning frame: Install Pi, wire it to one non-Anthropic model, build a small custom extension with a cheaper model, and observe where it hits its limits so you can hand only the final fix to a more powerful model.
- Why this matters: New playlist item from Cole Medin; queued for transcript-backed review, topic mapping, and a practical learning artifact.
Transcript anchors from this exact video:
- 0:14 / Evidence 1: "other coding agents right now. And so, the idea behind Pi is it is a minimal coding agent. It's made to be a coding agent that you build on top of instead of taking a really massive bloated..."
- 7:00 / Evidence 2: "great because it really is this minimal agent that allows us to use any model we could possibly dream of. Like we can use Kimmy K 2.6 for example, like I'll use in the stream today. But then..."
- 26:40 / Evidence 3: "factory. Curious how pi could help. So the the reason pi could be really helpful for the dark factory is because I have a very specific workflow for the way that I have the coding agent evolve the..."
- 28:12 / Evidence 4: "Okay. Um, Jcode. Sure. So I they Okay. So you can see that Pi is the second best, but apparently J-code is better with local embeddings off. I don't care that much about speed. Like here, here's the..."
- 49:01 / Evidence 5: "and rules and things like that, but it's like fundamentally even the coding agent itself, how it operates with the core loop, the core while loop of the agent. Like even that you can change. You can tweak..."
- 71:51 / Evidence 6: "really neat. The Codeex app specifically how you can manage your different projects and parallel agents and like the updates from there. It's kind of like the agent view that Claude Code released where you can see all..."
- 87:25 / Evidence 7: "extension so that the output of like the the last output from the workflow needs to go directly back into the the PI session I thought that's what you already had set up, but obviously not. It's very..."
Video-aware target:
- Prompt lane: Local model/runtime
- Mechanism to extract: Identify why the local setup works or fails for this specific agent task, not whether local models are generally good.
- Artifact to produce: A local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule.
- Artifact must include: hardware; runtime; model/quantization; endpoint; agent integration; benchmark/fallback
Your task:
1. Use the transcript anchors above as the primary source packet. If you add outside context, label it clearly as outside context and keep it secondary.
2. Create a source-check table with columns: timestamp, claim, transcript support, what the demo proves, confidence, and what still needs verification.
3. Extract the actual teachable mechanism from the video: Identify why the local setup works or fails for this specific agent task, not whether local models are generally good. Do not invent claims that are not supported by the title, lesson frame, or transcript anchors.
4. Build a reusable learning artifact: A local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule.
5. Include:
- a plain-English definition of the core idea
- a diagram or structured model using this sequence: Task -> Hardware -> Model/quantization -> Runtime endpoint -> Agent tool loop -> Benchmark task -> Fallback
- answers to these source questions: What machine/runtime is shown? | What task exposes the model limit? | What setup change improves the loop?
- 3 concrete examples that apply the video idea to real agentic work, such as Ollama or LM Studio coding endpoint; MLX Apple Silicon runner; DGX-backed Hermes session
- 2 failure modes the video helps prevent, chosen from the transcript evidence and these likely risks: using a local model like ChatGPT; ignoring latency/context limits; no benchmark task
- a checklist for the next real workflow, focused on: task fit, runtime setup, latency/context, tool loop, fallback
- one practical exercise with a clear done signal: Choose one real coding task and specify the pass/fail benchmark for a local model.
6. Add a "learning transfer" section: what changes in my workflow tomorrow if I actually learned this?
7. Add a "source check" section that cites which transcript anchor supports each major takeaway.
Quality bar:
- Make this specific to "Pi is INCREDIBLE - Building a Custom Coding Agent Live", not a generic Agent Architecture essay.
- Tie each harness element to a transcript anchor that names a tool, state boundary, permission, model behavior, or verification step.
- Prefer operational examples, failure modes, and reusable artifacts over broad definitions.
- Call out uncertainty instead of smoothing over weak evidence.
- Avoid these generic drifts: local-model ideology; hardware specs without workflow fit; benchmarks unrelated to the actual task.
- If evidence is weak or missing, stop and say what transcript segment or timestamp needs review instead of guessing.
- Finish with a concise artifact I could paste into my learning app.
Misconceptions
What to stop believing.
A better model automatically makes a better agent.
The model matters, but harness design determines whether the system can act safely and repeatably.
More tools always help.
Every tool increases surface area. Strong agents have the right tools with clear permissions.
Memory means saving everything.
Useful memory is compressed, curated, and tied to future decisions.
Practice studio
Learning only counts when you make something.
01
Transcript evidence map
Separate what the video actually says from what you already believe about the topic.
3 source-backed takeaways with timestamps, confidence, and a transfer note.02
One useful artifact
Apply the video to a real workflow and produce a local model/runtime fit sheet with hardware constraints, model choice, endpoint setup, task benchmark, and fallback rule..
A reusable artifact with a done signal and one verification step.03
Local model/runtime teach-back card
Explain the local model/runtime mechanism to someone who has not watched the video yet.
A 90-second explanation, one diagram, one example, and one misconception to avoid.
Recall check
Answer first, then reveal — without rewatching.
What is the design philosophy of Pi, and which model source is off-limits?
What issue did Cole hit with Pi's extensions during the stream?
What did building the Archon extension with Kimmy K2.6 reveal about cheap models?
Source shelf
Use the video as a doorway, then verify with primary sources.
