This was the easiest call and many like me made it in March among all of the anti-MCP wave of influencers[0][1] claiming it dead (many, many prominent folks in tech including Garry Tan).
It was so obvious why MCP would be needed once AI engineering and uptake moved beyond the solo developer and single harness stack of "what works for Me" versus "what works for My Team", particularly in an enterprise context. The key mistake people made was thinking in terms of their own workflows and own local stacks instead of a team's workflow and a team's operational stack. There was also an ignorance of MCP's stateless HTTP mode (yes, it was already a thing in March) versus local `stdio`.
I appreciate their reluctance towards MCP, but /something/ is better than nothing.
It’s suboptimal for the reasons the author outlines: but so is USB-C. So is NVME, so is HDMI.
We use these hugely successful technologies in spite of their flaws because they’re widely compatible and easy for the end user.
That’s why MCP is everywhere. It might not be performant, robust and uniform but it WILL get better over time.
And I’d much rather have the broad MCP ecosystem that we have now than seven or eight different “optimal” ways of plugging in an LLM to something useful.
Arguably, adopting ACP might even help Pi’s case, in that it could escape the terminal interface into one of many wrapper GUIs. TUIs inherently tend to limit your userbase to those who know what a terminal is… And given OpenAI’s latest product announcement [0] in response to Meta’s, the trend seems to be to attempt an expansion of customers to the general population, away from just developers and maybe “business” people.
Then again, I don’t even know if general adoption is what Pi/Earendil is going for.
I feel the same way about needing support for sub-agents, those feel pretty foundational to me.
I suspect that a smart model driving multiple dumber models for work and then using sub-agents with the same smart model for adversarial review will be a pretty common pattern.
Personally, I got a bit confused about Pi having most of that stuff as plugins since I remember how much of a mess Eclipse was where so much was just loosely fitting together plugins and just went with OpenCode since it covers most of my needs out of the box. Guess that might also be a sign of me getting older, because my IDEs and desktop environments are all closer to stock too.
Hah, it's the complete opposite for me :D. In Claude Code I disabled all sub-agents stuff, disabled nearly all tools but Bash, Edit, Write and WebSearch and replaced WebFetch with my own tool that doesn't summarize anything because the results were always worse with sub-agents, they always lack the necessary context and weaker models summarize bad. I also replaced the system prompt with my own that cuts a LOT of tokens, agents don't need a 10k+ system prompt anymore.
That's interesting! You don't have cases where the main session has important planning stuff but the actual work to execute has so much crap in it that context compaction will probably dig into the important plan stuff too much and make it too lossy? Also what about the cache read costs for longer context sizes?
Using sub-agents for example also lets me decrease the default context size in Claude Code instead of running at the full 1M like:
/autocompact 420k
or deal with Codex's 258k tokens (seriously quite tiny by modern standards).
running local models, now with Qwen3.8-Flash-Next, they have 256k, but when they get up there their speed is just too slow. So i've taken https://github.com/Tarquinen/opencode-dynamic-context-prunin... and started improving it. It already worked well to get a lot of mileage out of just taking tool calls, code modifications, etc, and dumping them in favor of a summary.
But they'd still inevitably get to long in the tooth, and context poisoning meant they'd just eventually not be able to stay in the preferred context size, which for me is 64k-128k. So, I extended it with an eviction command and required a ratio. So instead of a summary of work, it now just places a waypoint. The waypoint basically means the context has a semi-coherent context but without all the baggage.
I'm on like day 3 of a single session with 3m tokens removed and still in the sweet spot. So it evicts to beneath the lower limit, compresses to the upper limit, then evicts again.
It's amazing how resilient it is if you give it a good plan. The work flow has basically been:
1. Write up an implementation document for some new set of features.
2. Rewrite the implementation as a TDD document
3. Set it to work.
The only thing I haven't figured out is it likes to stop when it hits the finish line of the subparts, but likely we're going to end up with the master of puppets monitoring these things and just set them to evaluating what they've done.
> You don't have cases where the main session has important planning stuff but the actual work to execute has so much crap in it that context compaction will probably dig into the important plan stuff too much and make it too lossy?
For that is it not better to have separate sessions for planning stuff and doing actual work? Pi is super flexible with session management, and a lot of that can be automated by its extension system.
That sounds like you just NIH'd mr Zechner's Pi Coding Agent. Those were basically its founding design: yolo-mode security, simple design, minimalist system prompts, plug-in based for anything fancy (even sub-agents and web).
I did the same, also, the fact the tools evolve so fast, I dont want to waste time on a particular one while it might be obsolete next week. So either it works now, other I pick something else.
Exceptionally well is my takeaway. It’s the only harness I am using these days. I was previously using pi and codex mostly, but also the ones built into editors like zed, vscode, and the jetbrains IDEs.
On top of that, somewhat unrelated I’ll agree but still, it has support for vim keybindings
Pi vs omp is hotly debated within my friend group. It has most things you could want, ready out of the box, but also a lot of things you'd never want and it's constantly 5% broken. Some people love that trade, others don't.
Glad I'm not the only want to find it a bit janky/broken at times. They seem to constantly be pushing updates which is nice, but I treat it mostly as a black box.
Anecdotally, I find the auto compaction (or what I assume is happening when the context magically drops) to be hit or miss. I do like how easy it is to use my work cursor sub and business chat gpt at the same time. Then I use nearly free cursor models for dumb shit and Sol for real problems.
I did create some extensions where it spawns sub agents for specific tasks, especially when I want to keep the context clean or when I really want to offload a piece of work to a cheaper model. And for that I have a high degree of control over, I know which model is being used for each subtask.
I find Claude Code too unwieldy for my tastes. Pi's philosophy of being very light on features nut highly flexible for customization, clicked very well for the way I work.
Some things are impossible to just tack on or work around though, like MCP, while other things, can be done by just composing stuff.
Like sub-agents, you could just instruct pi/any harness with a user prompt/system prompt to start new invocations of itself, if you share what the exact command is, and pi or any other harness will do their own poor man's version of sub-agent via standard unix programs.
I had no idea pi didn't support MCP! I'm a new user, I just started messing around with it. I was getting my tooling up and running and tried to get one of my database MCPs working (Which, in retrospect, seemed a little painful - but I guess I was under the assumption that it was my responsibility to build + maintain those connections).
Another retrospect note, "No MCP" appears to be the first icon on their front page - not sure how I missed that.
... and after using it for a while you might end up forgetting most of the extra stuff you thought you needed in the first place. That's what happened to me and I've never looked back and still am a happy Pi user (https://a.l3x.in/ai if you're curious)
I'm still confused about what this codemode is. Models have been trained to chain bash and other typical unix tools well. They're so good at that to an uncanny level. Why do we want to not utilize this ability? Is it just a permission management issue in case you don't want the model to use shell directly?
We will write about it. The best way to think about it is that codemode solves a different problem than bash in that bash is a way for the agent to run a particular tool: running bash.
Codemode is a way for the LLM to orchestrate harness level tools. The reason this happening now, is because the models by the labs are increasingly trained on this. Codex for instance in responses lite requires codemode to even perform parallel tool calling.
As many others have commented: good, I too am an mcp hater.
However- in my testing, mcp is really quite fast, and its pretty much free at this point- with frontier models. Context rot is, from what ive tested, not as much of a concern now. I genuinely was not able to hillclimb skills/extensions to beat out the speed of mcp in some cases I've been testing.
I'm a bit bumped when I first saw pi is moving from bash to codemode and also adding MCP, since I thought that loses the purity and simplicity of the "use bash for everything" philosophy. However, after reading more about it, I realized codemode is just a slightly enhanced version of bash: more complex, sure, but likely more robust, secure, and efficient. For those who, like me, don't get the point of this change, here is how I understand it.
The first tool execution runtime in harnesses are direct tool calls with JSON or XML, such as the Read and Edit tools. As an escape hatch, we have Bash tool that allows arbitrary code execution on the host running the agent. The downsides of using bash (on the host) as the main tool execution runtime are:
- Syntax and obvious errors only surface at runtime
- Unergonomic orchestration of parallel and background tasks
- Verbose command output cluttering context
- Dependent on the host environment, packages versions, etc.
- No security measures by default.
To me the last point is the biggest inherent weakness, usually mitigated by creating a dedicated unprivileged user or running bash in a sandbox.
Note that direct tool calling is kind of the polar opposite on these points: syntax errors are caught early, orchestration can be done with some wrapping tools, command output is controlled, and most importantly they are more sandboxed. On the flip side, they obviously have way less power, necessitating Bash tool in the first place.
Codemode is the middle ground between these two extremes. It actually can be derived simply by one idea: what if we replace Bash by another language that can be checked for obvious errors, i.e. type checked?
Everything else falls out from there:
- Any language would do, but I think TypeScript fits the balance between safety, speed, conciseness, and popularity in training data.
- If we use TypeScript, might as well run it in a sandbox as JS runtimes have been designed with this in mind for 20 years
- Orchestration comes for free from the JS runtime. It's not more powerful, just more ergonomic.
- Since the tools are controlled by the harness and not dependent on the host, cloud agent becomes easier.
- With this in place, MCP are not very different from a tool provided to this sandboxed runtime.
Overall I find the benefits compelling enough, but we'll see if the heavily-RLed models these days will use it effectively.
The core idea of Codemode that I understood is - The script doesn't actually write any code itself; rather, it acts as a workflow automation and program management tool. It pulls data from an issue tracker, delegates the analysis to a model, and synthesizes the results to help manage Pi's development priorities.
> And while we could have just wired up the metadata to enable better MCP extensions, we also think that MCP with Codemode solves quite a few of the issues that it traditionally had.
There's just something that bothers me about this. Normally if LLMs want to compose multiple operations, they have the perfect tool for this: bash, or whatever other OS shell is available. It's why I was always confused by Codemode-type constructs for direct chaining of tool calls; see also the way highly-RL'd modern models will fall back to sed or python for complex file edits.
It seems like Codemode is raised here as the perfect tool for chaining or composing MCPs, but isn't that backwards? LLMs are already given the perfect tool for that, and the problem is that MCPs aren't exposed to that tool.
I happen to think codemode is useful, but not the full answer. I have a long and skewed history with chaining things in MCP and built a dozen or so enterprise ones (see https://taoofmac.com/space/blog/2026/04/29/2341 for notes) and it all falls back into the trade-off between agent scope/context and tool coverage: If you are using a coding agent it will have no trouble sorting out any tool regardless of how many are exposed (it's just a matter of either progressive tool disclosure or good tool metadata, since the coding agent will just go at it and expend whatever tokens are needed), whereas in a "normal", limited, scoped agent that has only a few things it needs to do (like handling a ticketing system) codemode is pretty much overkill.
Pi is primarily a coding agent, so yeah, code mode makes sense, but I've found that better MCP design saves everyone a lot of trouble and would also probably have improved the thing's reputation overall (I personally am not fond of the line protocol, would rather have protobuf and more typing, but it is what it is).
> Normally if LLMs want to compose multiple operations, they have the perfect tool for this: bash, or whatever other OS shell is available.
I many scenarios, e.g. running the harness server-side, as is the case for chat interfaces, you don't really want to expose OS shell access as that opens up a huge security attack surface.
> I many scenarios, e.g. running the harness server-side, as is the case for chat interfaces, you don't really want to expose OS shell access as that opens up a huge security attack surface.
It does, but a restricted user account mitigates the large majority of those issues. A sandbox mitigates even more.
The number of remaining exploits left is probably going to be the same as the number in the harness. More, in fact, as many of them have no human review anyway.
Cloud products based orchestrations with proper security mechanisms configured, don't have shell access and should only communicate over proper network mechanisms.
Rootless immutable containers without shell access, or SaaS products from multiple vendors with WebAPIs as the only touch point.
the conversation seems to dwell on things you could substitute Bash for but the real need stems from completely opaque systems that nothing can reach but which are now getting MCP support. This is where being left out of having MCP support will hurt. I'm still quite happy to let all the harnesses compose bash commands to their hearts content (inside their sandboxes ...)
You can turn those tools off. In fact, that is what I am doing right now in https://github.com/rcarmo/piclaw until I am positive the new MCP stuff has full parity with the MCP adapter I've been shipping for the past six months or so.
I'm actually pretty happy that they did it, since 90% of what I have to integrate in enterprises is MCP-driven (it's a security and auth boundary that has become pretty much mandatory for any third-party agents wanting to reach into corporate data) and this lets me use Pi directly. Am just being cautious about the first version, because, well... it's a first version, and I like my tools stable.
(I actually played around with the idea of using QuickJS myself for codemode, but since I rely on Bun that gives me the ability to use other things... never got around to do it though.)
Ok! I trust that you and the maintainer will steward the project properly. It's just that I really like Pi as is and am a natural worrier. I'm also not sold on Jev(-likes), so that reasoning rung a bit hollow to me.
This is somewhat similar to HuggingFace smolagents where the model writes code that calls tools, instead of emiting json to describe the tool call per turn. Here Codemode is one tool that the model calls when it needs to compose many tool calls, especially MCP ones. Is what i understand of this.
Good. I too I'm not a fan of MCPs, but these days I do find them useful. In Claude Code I connected to my company's MCP which made Claude Code infinitely more useful for everyday work stuff
Have a go at https://github.com/rcarmo/memento, I would appreciate Claude testers since I mostly use Codex. Just trying it and filing an issue about what doesn't work would be great...
Honestly, the provided argument for it is rather weak. They are basically adding a way of running scripts that are contained within harness to execute harness's own tools (that's the Codemode). A coding agent can already compose any arbitrary logic by invoking shell scripts (or python scripts, or node scripts), etc - so this is just entirely unnecessary in the core, from my perspective.
If you feel that Pi has been drifting away from its original vision, try hax (https://usehax.dev/) - you might like it.
Theres transport and context management. A cross agent status cache with semaphore over a testing harness driving a browser that can rewind and retry is much easier for agents to drive from mcp than selenium or whatever is fashionable these days. Didn't replace unit testing per se but to rca and fix it's a much better tool.
While we are waiting on that to become stabilized, we implemented a inspired/co-evolved way to do that in our tool[0], where you mark individual fields in the request/response schema as being file payloads, so that file exchange can be properly orchestrated by the harness and doesn't pollute the context. We just do inline base64 uploads of the required payloads, which in practice we've seen to work quite will until ~100MB files (which is otherwise also the size limit we usually recommend for file processed).
It's annoying that it's not stabilized yet, but for most bigger customers we've seen, they implement 80% of the MCP servers they connect in-house, so doing adjustments to the tool surface, and metadata has been less of a pain for them than we expected.
I have had to hack a couple of workarounds in https://github.com/rcarmo/memento to do uploads, and there's a draft going around, but the general practice in enterprise MCPs seems to be to do it "out of band" and have MCP tools to hand-over storage handles/URLs so the MCP server can do the imports itself "safely".
Most people using pi probably know. MCP is “model context protocol”, a protocol by which models can connect to apis and services and conversely a way to expose those apis and services so they can be used by llms and agents. https://modelcontextprotocol.io/docs/2026-07-28/getting-star...
How on earth does the article no address the title? Literally the first paragraph basically "spoils" the entire article and you have your answer, then you can continue reading for more justification of why it was like X before but now it's like Y.
Author of the post here: I don't think it's a good idea to put your clanker to a PR and then try to explain it. That's because you are then reading a derivative work of a derivative work instead of going to the source.
If we fail to explain it, then we need to do a better job explaining it :)
Yeah, since pi is my first agent harness I barely understand what MCP is beyond an API standard. When the essay started talking about codemode without any introduction I was having trouble following it.
It was so obvious why MCP would be needed once AI engineering and uptake moved beyond the solo developer and single harness stack of "what works for Me" versus "what works for My Team", particularly in an enterprise context. The key mistake people made was thinking in terms of their own workflows and own local stacks instead of a team's workflow and a team's operational stack. There was also an ignorance of MCP's stateless HTTP mode (yes, it was already a thing in March) versus local `stdio`.
[0] https://news.ycombinator.com/item?id=47380270
[1] https://chrlschn.dev/blog/2026/03/mcp-is-dead-long-live-mcp/
It’s suboptimal for the reasons the author outlines: but so is USB-C. So is NVME, so is HDMI.
We use these hugely successful technologies in spite of their flaws because they’re widely compatible and easy for the end user.
That’s why MCP is everywhere. It might not be performant, robust and uniform but it WILL get better over time.
And I’d much rather have the broad MCP ecosystem that we have now than seven or eight different “optimal” ways of plugging in an LLM to something useful.
I hope they'll do the same and eventually add native support for ACP (https://agentclientprotocol.com/get-started/introduction) which, on the contrary, I use quite.
Then again, I don’t even know if general adoption is what Pi/Earendil is going for.
[0] https://news.ycombinator.com/item?id=49896604
I suspect that a smart model driving multiple dumber models for work and then using sub-agents with the same smart model for adversarial review will be a pretty common pattern.
Personally, I got a bit confused about Pi having most of that stuff as plugins since I remember how much of a mess Eclipse was where so much was just loosely fitting together plugins and just went with OpenCode since it covers most of my needs out of the box. Guess that might also be a sign of me getting older, because my IDEs and desktop environments are all closer to stock too.
Using sub-agents for example also lets me decrease the default context size in Claude Code instead of running at the full 1M like:
or deal with Codex's 258k tokens (seriously quite tiny by modern standards).Same idea with something like OpenCode, there I even configured custom agents for review: https://opencode.ai/docs/agents/
But they'd still inevitably get to long in the tooth, and context poisoning meant they'd just eventually not be able to stay in the preferred context size, which for me is 64k-128k. So, I extended it with an eviction command and required a ratio. So instead of a summary of work, it now just places a waypoint. The waypoint basically means the context has a semi-coherent context but without all the baggage.
I'm on like day 3 of a single session with 3m tokens removed and still in the sweet spot. So it evicts to beneath the lower limit, compresses to the upper limit, then evicts again.
It's amazing how resilient it is if you give it a good plan. The work flow has basically been:
1. Write up an implementation document for some new set of features.
2. Rewrite the implementation as a TDD document
3. Set it to work.
The only thing I haven't figured out is it likes to stop when it hits the finish line of the subparts, but likely we're going to end up with the master of puppets monitoring these things and just set them to evaluating what they've done.
Doesn't that destroy the cache? I find that caching significantly sped up my Qwen, especially on said larger contexts.
For that is it not better to have separate sessions for planning stuff and doing actual work? Pi is super flexible with session management, and a lot of that can be automated by its extension system.
https://github.com/can1357/oh-my-pi
I haven't tried it much though, can't vouch how well it works.
On top of that, somewhat unrelated I’ll agree but still, it has support for vim keybindings
Anecdotally, I find the auto compaction (or what I assume is happening when the context magically drops) to be hit or miss. I do like how easy it is to use my work cursor sub and business chat gpt at the same time. Then I use nearly free cursor models for dumb shit and Sol for real problems.
I did create some extensions where it spawns sub agents for specific tasks, especially when I want to keep the context clean or when I really want to offload a piece of work to a cheaper model. And for that I have a high degree of control over, I know which model is being used for each subtask.
I find Claude Code too unwieldy for my tastes. Pi's philosophy of being very light on features nut highly flexible for customization, clicked very well for the way I work.
Like sub-agents, you could just instruct pi/any harness with a user prompt/system prompt to start new invocations of itself, if you share what the exact command is, and pi or any other harness will do their own poor man's version of sub-agent via standard unix programs.
The idea of using jev as a cheaper faster subagent for specific use cases is interesting. Will have to experiment with that!
Another retrospect note, "No MCP" appears to be the first icon on their front page - not sure how I missed that.
Imagine my surprise reading this!
Codemode is a way for the LLM to orchestrate harness level tools. The reason this happening now, is because the models by the labs are increasingly trained on this. Codex for instance in responses lite requires codemode to even perform parallel tool calling.
However- in my testing, mcp is really quite fast, and its pretty much free at this point- with frontier models. Context rot is, from what ive tested, not as much of a concern now. I genuinely was not able to hillclimb skills/extensions to beat out the speed of mcp in some cases I've been testing.
The first tool execution runtime in harnesses are direct tool calls with JSON or XML, such as the Read and Edit tools. As an escape hatch, we have Bash tool that allows arbitrary code execution on the host running the agent. The downsides of using bash (on the host) as the main tool execution runtime are:
- Syntax and obvious errors only surface at runtime
- Unergonomic orchestration of parallel and background tasks
- Verbose command output cluttering context
- Dependent on the host environment, packages versions, etc.
- No security measures by default.
To me the last point is the biggest inherent weakness, usually mitigated by creating a dedicated unprivileged user or running bash in a sandbox.
Note that direct tool calling is kind of the polar opposite on these points: syntax errors are caught early, orchestration can be done with some wrapping tools, command output is controlled, and most importantly they are more sandboxed. On the flip side, they obviously have way less power, necessitating Bash tool in the first place.
Codemode is the middle ground between these two extremes. It actually can be derived simply by one idea: what if we replace Bash by another language that can be checked for obvious errors, i.e. type checked?
Everything else falls out from there:
- Any language would do, but I think TypeScript fits the balance between safety, speed, conciseness, and popularity in training data.
- If we use TypeScript, might as well run it in a sandbox as JS runtimes have been designed with this in mind for 20 years
- Orchestration comes for free from the JS runtime. It's not more powerful, just more ergonomic.
- Since the tools are controlled by the harness and not dependent on the host, cloud agent becomes easier.
- With this in place, MCP are not very different from a tool provided to this sandboxed runtime.
Overall I find the benefits compelling enough, but we'll see if the heavily-RLed models these days will use it effectively.
Codemode isn't replacing MCP; it's fixing MCP's biggest flaw—its lack of composability.
There's just something that bothers me about this. Normally if LLMs want to compose multiple operations, they have the perfect tool for this: bash, or whatever other OS shell is available. It's why I was always confused by Codemode-type constructs for direct chaining of tool calls; see also the way highly-RL'd modern models will fall back to sed or python for complex file edits.
It seems like Codemode is raised here as the perfect tool for chaining or composing MCPs, but isn't that backwards? LLMs are already given the perfect tool for that, and the problem is that MCPs aren't exposed to that tool.
Pi is primarily a coding agent, so yeah, code mode makes sense, but I've found that better MCP design saves everyone a lot of trouble and would also probably have improved the thing's reputation overall (I personally am not fond of the line protocol, would rather have protobuf and more typing, but it is what it is).
I many scenarios, e.g. running the harness server-side, as is the case for chat interfaces, you don't really want to expose OS shell access as that opens up a huge security attack surface.
It does, but a restricted user account mitigates the large majority of those issues. A sandbox mitigates even more.
The number of remaining exploits left is probably going to be the same as the number in the harness. More, in fact, as many of them have no human review anyway.
That's been a trivially solved problem for decades.
Rootless immutable containers without shell access, or SaaS products from multiple vendors with WebAPIs as the only touch point.
But pi doesn't have security. Pi is full yolo, it's on the user to run it in an environment that minimizes the blast radius if the LLM goes haywire.
So I'm not sure what the plan is here. Will pi support running certain tools like bash as a different OS user than owner of the pi process?
I'm actually pretty happy that they did it, since 90% of what I have to integrate in enterprises is MCP-driven (it's a security and auth boundary that has become pretty much mandatory for any third-party agents wanting to reach into corporate data) and this lets me use Pi directly. Am just being cautious about the first version, because, well... it's a first version, and I like my tools stable.
(I actually played around with the idea of using QuickJS myself for codemode, but since I rely on Bun that gives me the ability to use other things... never got around to do it though.)
If you feel that Pi has been drifting away from its original vision, try hax (https://usehax.dev/) - you might like it.
Just generate CLI tools, with docs, from MCP servers on demand.
Swagger was the new old name for the concept
While we are waiting on that to become stabilized, we implemented a inspired/co-evolved way to do that in our tool[0], where you mark individual fields in the request/response schema as being file payloads, so that file exchange can be properly orchestrated by the harness and doesn't pollute the context. We just do inline base64 uploads of the required payloads, which in practice we've seen to work quite will until ~100MB files (which is otherwise also the size limit we usually recommend for file processed).
It's annoying that it's not stabilized yet, but for most bigger customers we've seen, they implement 80% of the MCP servers they connect in-house, so doing adjustments to the tool surface, and metadata has been less of a pain for them than we expected.
[0]: https://erato.chat/docs/features/mcp_servers#file-support
I find this approach is easier to debug and I can also use the tool myself to ensure it's working well.
And you didn’t remember that when you said no to MCP?
No, no MCP for now?
If you are a Pi user it may be better to just ask your agent to explain https://github.com/earendil-works/pi/pull/10040
If we fail to explain it, then we need to do a better job explaining it :)