Rendered at 13:48:26 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
kierangill 15 hours ago [-]
Instead of compaction, has anyone seen a successful implementation of pruning? That is, the agent looks at the conversation history and removes any low-value messages.
For example, sometimes context will be taken up by a side tangent, tool call outputs, or low-value codebase exploration.
Much of the time, I prefer to preserve the history of my conversation instead of summarizing it. I find summarized conversations lead to more frustrating future chats because the LLM misses intent and or context. (Or, the presence of paragraphs and paragraphs of LLM output makes the next token predictor dumber? Unsure.)
I’m still playing with it, but it essentially has the model define a region, and mark it with begin_task, and end_task. End_task also requires a fairly hefty summery of the entire region.
It then pulls that entire region and replaces it with the summary. If the model wants it can actually look into the output that was pruned from the conversation, but in practice the summaries have been good enough it hasn’t needed to.
I’ve been using it with 5.6 sol, and even really long sessions have like 6% context used at the end.
pjm331 14 hours ago [-]
I think there are a lot of strategies that will open up when costs come down but right now you take a hit on cache rate and thus costs every time you do anything other than wait until the last minute to compact
charcircuit 7 hours ago [-]
You can do things like throwing away or summarizing information from the most recent calls so you fall back to a recent checkpoint that is still cached
toasty228 1 hours ago [-]
In oh-my-pi there is the /shake command that gets rids of tool results and some other things
mikepurvis 13 hours ago [-]
I'm very interested in this too. I feel like when my Claude session compacts I immediately have to re-alert it to critical aspects of the task, but it should be possible even to have a secondary low-skill agent crawl over the whole context window semi-continuously and nominate pieces for removal or summarization. Or at a higher level, have something able to decompose, like hey I realised we're actually working on three distinct aspects of this problem, let's split the context into those three and I'll just retain a high level summary of what's in the other ones, enough to know if I should pass off control to one of the others for a particular ask.
I feel like this is kind of subagents, but it doesn't have the clear task -> work -> return result -> die model, rather it's more closely aligned to how I myself multitask, maintaining several long-lived threads of execution, and having to intentionally "forget" about the other ones when it's time to lock in and do real work on the most important one.
trueno 3 hours ago [-]
yeah ive read so many takes on how people survive compaction amnesia ive wired up my own "protocols" that are okay, have used things like beads, am okay burning a buttload of context right at the beginning of a post-compaction just to get up to speed and oriented.. but i still end up with a decent amount of sprawl and lately the writing style of claude has just gotten so dense and nonsensical that my eyes are starting to glaze over.
this seems to be the fate with all the big grand 'god' projects people are doing with AI. super epic omega knowledge bases, super epic omega personal platforms, etc. I don't think enough people admit that whatever they're working on has evolved into something they don't fully understand.
Schlagbohrer 4 hours ago [-]
I have run in to the same thing. Most recently I told my pi-agent to fork a git repo since our work had morphed into two different tasks that shouldn't be packed into the same context and workflow. Git fork is what I will use as a tool for this division in the future.
d4rkp4ttern 12 hours ago [-]
I have a few variations of pruning (trimming all long tool outputs, agent-directed trimming, etc) in my aichat collection of session tools:
OMP has this, it's called "/shake" where bloat from tool calls and such is turned into an artifact ref instead of remaining in the context.
There are other pruning or compaction strategies you can configure too.
You can also use "/btw" to have side tangent conversations with the current context but with the benefit that it will not bloat your current context after the tangent is done.
I wish other harnesses were prioritising this sort of flexibility with context management, but they're likely more incentivised on optimising their cache rates
> You can also use "/btw" to have side tangent conversations with the current context
Ah, something I needed, but did not know I needed. Comes up ALL the time. I want to know something that is well off the task, but it's the best time to ask due to the specifically loaded context.
jsw97 12 hours ago [-]
I built a harness for my own use that allows the agent to fork its own history. So for example it can “compact from” a specific item, replacing a branch with a summary; or do an “excursion”, basically a temporary branch, which is like a subagent but inherits context.
Sounds cool and it does make sensible decisions optically but I haven’t been able to prove that it is meaningfully better than normal compaction. Building harnesses that do interesting things is a lot easier than building more effective harnesses, I guess.
WhyNotHugo 9 hours ago [-]
Potentially: remove thinking blocks, and keep the rest. At least this would ensure that the entire context of the conversation is still there, and anything said isn't lost.
Having a second model also iterate the resulting messages and remove low-value tool calls could also be interesting. Especially failed calls which add no value.
Wouldn’t that invalidate kv cache and you wouldn’t benefit from much cheaper cache read?
Dayshine 8 hours ago [-]
I'm frequently idle for five minutes while I review, at which point the cache has expired
stpedgwdgfhgdd 7 hours ago [-]
[dead]
tosh 6 hours ago [-]
when you look at the compaction prompt: in a sense it is doing that pruning but the llm decides what to prune
zmgsabst 8 hours ago [-]
I do this in my own harness, where a context only exports certain messages when you step out of it — eg, my request for an essay and its final output, while dropping everything from the first draft through various intermediary edits.
That naturally trims a lot of context while removing mistakes from the context to prevent poisoning (ie, every draft but the final negatively contributes in some way — that’s why we edited them).
Because juggler has a nested-thread architecture, I had a great ah-ha! moment when I realised that with sub-threads, compaction becomes almost free: you just take the list of items in the conversation, move them all into a new sub-thread, then allow that sub-thread to summarise itself (this summary is what the parent thread sees). Hey-presto, your parent thread is now compacted, and the sub-thread contains all the old messages, so you can browse them if you need to, or just undo the whole operation if you change you mind. This struck me as super-elegant, and I was chuffed that it works.
However.. What I find myself actually doing now is just using the "new conversation" tool to hand over to a new conversation - I just tell the LLM "create a new conversation, and tell it everything it needs to continue this task". It gets you to the same place, but generally I'm not interested in the history, and having a fresh conversation with a new name usually feels neater.
Imanari 2 hours ago [-]
Great thread, I was just thinking about compaction. My current line of thought is that compaction/pruning/ctx management in general should be something ongoing and maybe recursive. For example:
User:'How is auth implemented?'
->
[thinking]
[codebase exploration with [thinking] in between, 10 file reads, 3 of which were "wrong"]
[thinking]
->
agent_response
This little exchange contains a WHAT (how auth actually is implemented) and a HOW (where that info is and how to retrieve it). Maybe this question was part of a larger task. I think that whole exchange could be summarised before it enters context, kind of like what happens with subagents. The main thread would then consist mostly of [summaries]. Eventually the context will fill up anyway and we would summarise those summaries again. Alternatively one could maintain a [master_summary], kind of like an internal state. So new [summaries] get integrated directly and the [master_summary] gets updated.
errantmind 10 hours ago [-]
In my experience, the best approach to compaction is to never get to the point where you need compaction and to generally stay below about 30% context window utilization. Even for long agentic workflows this can be accomplished for quite a while, much longer than most people might think.
Here's what I do for each of my sessions:
1. For asides, off-topic work, or repetitive work that has already been done in the session, branch backwards (with /tree) and summarize.
2. If I've exceeded 30% or the 'price-doubling' multi-tier pricing, prune (my custom extension).
3. If I've already pruned and I'm still close to 30%, 'prune all' (more extensive prune).
Definition:
'/prune': Removes ~50% context on a fresh session (not previously pruned)
- Keeps: User messages, normal assistant prose, commands/status markers, extension receipts, model settings, and a plain-text receipt for each tool call.
- Removes: Thinking, signatures, actual tool calls/results, tool output, images, compaction summaries, and other extensions’ state.
'/prune-extended': Removes ~80% context on a fresh session
- Keeps: User messages, normal assistant prose and conclusions, commands/status markers, extension receipts, and model settings.
- Removes: Thinking, signatures, all tool calls/results and output, images, compaction summaries, other extensions’ state, and any tool-activity receipts created by /prune.
Both create a new session and delete the old one after a successful switch.
Using these I can keep a session going for weeks (or longer), even with extensive use and almost all the important context is preserved while dumping the less important context. Neither command requires an LLM summarization so they execute quickly.
Imanari 3 hours ago [-]
I am surprised at 'removes actual tool calls/results, tool output'. Your approach with /prune seems to be 'keep the WHAT, remove the HOW (we got here)'. I would have thought that the HOW contains some useful signal.
hatthew 9 hours ago [-]
What is that 30% number based on? Surely that's a model specific limit, and is based more on the absolute token length, not percentage, right? I'm not sure it makes sense for e.g. Opus 0.2M and Opus 1M to both degrade at 30% of their respective context lengths.
aevv 8 hours ago [-]
https://www.producttalk.org/context-rot this article summarises several papers that have explored this, and it does seem to be related to both the absolute number of tokens and the % of the window
bvrmn 3 hours ago [-]
People nowdays have no shame and completely forget the art of plug.
novaRom 16 hours ago [-]
Compaction is painful if you run just one local LLM, the best way to avoid it is to keep context as small as possible.
One trick I find useful is to have one model with two KV caches running and while first cache has produced tokens, second cache immediately summarizes them during input tokens are being generated (tools time), then harness switches to the second KV cache which takes newly produced input tokens while KV in first cache is getting replaced with compacted summary tokens. This is a kind of ping pong, so we trade more space for less time. Still experimenting but it looks it works, and nice bonus it improves GPU utilization. Btw I have my own harness and model serving code, but it can be easily implemented in any other harness and model server.
kristianpaul 14 hours ago [-]
I tend to ask a line of questions to the LLM as first step to a point the context is enough for me. Then i ask for the work i need but then get back to previous context using /tree.
It's like asking the questions needed to get the answer you need from the LLM, can be either an answer for a plan/todo or a task to code/change something.
skeledrew 14 hours ago [-]
I think the way prompt caching works really discourages more creative compaction techniques. Like perhaps some kind of heuristic progressive compaction that replaces tool results and thinking traces after use with pointers could potentially keep the model smart for much longer, but that'd mean breaking cache every turn, and possibly even within a turn, seriously driving up cost.
damsta 16 hours ago [-]
I don't like any of current solutions when it comes to compaction. I'd love to have a way to say what exactly should be summarized, because most of the time I just need to compact some noisy MCP tool calls, test runs and things like that. Just let me pick what should be summarized and keep the rest as is.
flexagoon 16 hours ago [-]
You can do that in Pi!
> Extensions can intercept and customize both compaction and branch summarization
Just make an extension (or ask Pi to write an extension for itself) that intercepts compaction and leaves only what you want, or rewrites it in any other way. Should be just a few lines.
alansaber 15 hours ago [-]
Sounds like you might like subagents. Agent > subagent receives agent context (presumably cached)->tool call->compact/summarise->return to main agent
lightbendover 15 hours ago [-]
[dead]
boorang 10 hours ago [-]
I mean, not to be flippant but can't you just prompt the agent to write a file as you're getting closer to the compaction limit? I tend to just go to roughly 50-70% context utilization and then tell the agent to summarize the conversation and save it to a file, manually /clear, then say let's continue that last conversation. You can inspect the summary first and make any changes.
zavec 2 hours ago [-]
The way I've done it has been when there's a longer task, I give it a markdown document that has a plan with numbered steps, and then for each step start a fresh context window and tell it to update that doc as it goes with any decisions taken or deviations from the plan, or other context needed for future steps. If it gets close to the end of the context I tell it to summarize the current state and anything a fresh context would need to know.
fermuch 13 hours ago [-]
I think /handoff on pi (or at least oh my pi) is what you are looking for
OMP changed the default compaction to images! Kinda nuts to read about. Saves the generation cost of the traditional compaction step and writes the context as tiny text to an image, if I was following correctly.
bakies 11 hours ago [-]
What's helpful about that?
jubilanti 8 hours ago [-]
This is one of the most ridiculously awesome findings this year, that changed my mental model of how foundation models work.
It sounds crazy, but it's way cheaper and faster for multimodal models to read text on images than the equivalent text represented as tokens. Text tokens take a lot more bytes to represent internally than you would think. Give it images with rasterized text, and there is no OCR-like intermediate step that turns images into text tokens. It just goes straight into the multimodal context where it's all just linea algebra underneath. The cost is more lossiness than if you gave it the same tokens as text. Which is why you don't just render text prompts and files to read as images.
But compaction is inherently lossy, a summary 1-10% of the original length is going to lose things, by definition. So you keep the things you really need to get exactly right in text token summary, but put a lot more of what you would have just discarded into a PNG. Keeping the past context as text in an image is way less lossy than keeping a text summary of that context.
They peeked inside the layers and found the same neurons were usually activating in the middle layers when fed either text tokens or images of that same text. It just takes the model more time and kv cache to do it via text tokens.
But what if the model you're using doesn't have image processing capabilities?
jakswa 6 hours ago [-]
I dunno, I didn't read in-depth. Hopefully you don't gotta zoom in with human eyeballs.
Aeolun 8 hours ago [-]
My summarization creation functions over batches of 50 messages, and I don’t often lose important context any more. The loss comes from trying to stick a whole conversation in a single compaction request (at least in my case)
imgyuri 3 hours ago [-]
the way compaction works is pretty simple and highly relies on another model to do it, the only part the user has control of is when to do the compaction, which actually means when doing large amount of work in one shot, it should be planned from the beginning to be separated into works that can be reviewed, afterwards the model can compact
zahrevsky 16 hours ago [-]
Was expecting the article to go more in-depth.
Say, what happens when chain of summaries grows so long, that it still overflows context window. Is summarization runned over the summaries in the context window?
alansaber 15 hours ago [-]
What more depth is there to go to? Compaction is a single LLM call (practically) which can have some deterministic diffing/extraction baked in, or multiple LLM calls (generally wasteful). There's only 1 summary in the context window at one time. Every prompt goes [CONVERSATION_HISTORY] + input -> model turn. As soon as total context exceeds that it compacts, so there's no summarisation overflow (you can enable an agent to access past summarisations from past compactions, but the snake starts to eat it's own tail).
skeledrew 14 hours ago [-]
There is no chain, just a lossy summary that includes a summary of any previous summary along with the fresh messages.
storus 16 hours ago [-]
The advantage of running local stack is that you can do the compaction at the time of inference, i.e. some tool call runs out of context, you can just pause inference, purge/replace old tool calls with their summaries or just logs by operating directly over tokens on a GPU, rebuilding KV cache (one time prefill hit) and resuming the inference, easily being able to e.g. read 1000 markdowns, each 50k long, in a single LLM call. That's not possible with current agentic harnesses using LLM calls.
kennywinker 17 hours ago [-]
Compaction has been a pretty painful part of local llm usage. Scrapping the current context and parsing almosy 128k of context then generating something like 5-10k tokens - that can take quite a while when you’re working with 10t/s-45t/s (depending on the model).
I pretty much just start a new session whenever i fill the context.
xyzzy123 16 hours ago [-]
In my opinion this is one of the areas where GPUs provide a qualitatively different experience than unified memory boxes.
For an EPYC with a 5090 (no layers on CPU) vs an M3 max 128GB, qwen 3.6 27B at 128k context / 7k generation:
Cold: prefill + decode Hot (KV cached)
5090 40s + 2-3m = 3-4 min 2-3 min
M3 Max 128GB 14m + 8-10m = 22-25 min 8-10 min
This is for dense qwen (which I wouldn't run day to day on the mac) - in reality the mac is quite usable with MoEs but you definitely notice a difference.
pornel 14 hours ago [-]
I don't like that it throws away the whole KV cache when compacting. It costs a cache miss of the whole conversation length, and that's a waste of time and money.
LLMs are perfectly capable of summarising the conversation without a new system prompt.
meatmanek 13 hours ago [-]
Yeah, that seems suboptimal, unless you're using a much cheaper model to do compaction.
kristianpaul 14 hours ago [-]
How the summarizing of the conversation happens in an LLM?
pornel 10 hours ago [-]
Instead of prompting "Summarize: $SESSION" you submit "$SESSION Summarize what has been done so far".
This way you have a prefix cache hit on the session and don't pay any more than for a regular prompt.
pranayVarma0512 5 hours ago [-]
This seems to be a very intuitive take, but people often miss on the technicalities.
The prompt is assembled as tool schemas, then system instructions, then message history, and caching works on prefixes of that. So the cost of an edit is not its size, it is the size of everything behind it. Edit a tool definition and you have invalidated the system prompt and the whole conversation with it. Edit the tail and you pay for the tail.
I measured this by accident in a multi agent ablation where the only variable was whether the supervisor sent a fixed tool array or a per-task subset to the worker. Runs paying cache creation, out of 120 each: fixed 0 cache creation, per-task subset 58 cache creation. Under a prompt load that was $0.0382 per run against $0.0230, and on a clean context it reversed because there was no prefix worth caching.
Which is the argument for what Pi does here. A pointer is a tail edit. Rewriting is a head edit.
navs 10 hours ago [-]
Ampcode used a handoff feature for a while that I found genuinely useful [1] and then they removed it. Anecdotally, I felt it worked better than compaction.
Can someone recommend a Hermes alternative that is less token hungry? Pi did not work well for my use case.
aolsenjazz 27 minutes ago [-]
What about pi doesn't work? I've been enjoying dcode
skinfaxi 17 hours ago [-]
I found hermes to be really lightweight, though I am on a relatively older version and built a custom plugin to lazily load mcps (that's probably in hermes proper by now). Compared to kilo it seems to consume far fewer tokens.
pkulak 16 hours ago [-]
I keep my max context really small for personal assistant agents; they don't need it. Especially since compaction keeps anything important around anyway. I use 60k with Pi.
alfiedotwtf 4 hours ago [-]
I want to like Pi but compaction is why I had to go back to Open Code…
My problem is that when in a loop and it’s calling tools, it won’t check how close it is to the compacting limit until the whole loop returns to you for the next prompt. And so if you have a run that could go for hours, it’s a gamble if you’ll OOM or an interrupting compaction breaks context and stops the loop without continuing.
There’s a few extensions that all try to solve this problem, but I’ve found none actually work :(
brcmthrowaway 13 hours ago [-]
Is pi a drop in replacement for OpenCode / Claude Code? I cannot be bothered installing 50 million plugins.
josh_p 13 hours ago [-]
It is not.
Their approach is minimalism. If you need something pi doesn’t have, the recommendation is to ask pi about it and have it build an extension for you, or look for a plugin.
That said, try it. You may find you didn’t need everything in those ecosystems.
josh_p 13 hours ago [-]
I also find the SDK really valuable. Being able embed a truly minimal agent wherever you can run node/typescript is very powerful. Any model, any provider, full control over what it’s doing.
I expect Pi is mostly used with OpenAI plans, and OpenAI has a dedicated compaction endpoint you should probably be using with their models instead of a compaction prompt.
brandall10 15 hours ago [-]
There is a compaction routing plugin, as well as a general OAI compatibility plugin that simulates some of their harness features (ie. code_mode/batching tools) that also includes their compaction.
That said, from reading the pi subreddit I don't see people referencing these much.
ka_de 4 hours ago [-]
[flagged]
randomblock1 17 hours ago [-]
TLDR: It keeps ~20k tokens of recent conversations, then hands the rest of the conversation to another model with a special system & user prompt. This then fills out a template with relevant information.
Can also be the same model. The main thing is that it uses a separate, clean context window.
alansaber 15 hours ago [-]
Sounds just like opencode.
tizerluo 12 hours ago [-]
[flagged]
cyanydeez 17 hours ago [-]
Opencodes dynamic context pruning works by labeling tools and chat and the rest and the agent can collapse and expand summaries.
I get it into 1M+ routinely on local models with operations between 50k-85k
skeledrew 14 hours ago [-]
That dynamic pruning sounds bad for prompt caching though.
cyanydeez 13 hours ago [-]
sure, it's the price you pay, but you have to considered that context poisoning is basically a statistical certainty approaching 1 regardless of model size. As you grow the context, the likelihood of vague details getting conflated increases, so if you're _not_ cutting down the context you're increasing the probabilities of just basic random jitter in your logic and code.
The pruning does evict cache but not from the start, and only up till the last time you pruned it, as it's just adding messages on top.
but yes, it takes some additional tests and docs to keep it from just becoming hollowed out on tasks; I'd say about 10-20% of the time is just horribly loses what it's doing.
juleiie 16 hours ago [-]
[flagged]
miningape 15 hours ago [-]
Yeah, and where's Bombadil Inc.?
bellowsgulch 15 hours ago [-]
We must create new stories for good so that evil has something to steal! We're running out of unironic irony!
For example, sometimes context will be taken up by a side tangent, tool call outputs, or low-value codebase exploration.
Much of the time, I prefer to preserve the history of my conversation instead of summarizing it. I find summarized conversations lead to more frustrating future chats because the LLM misses intent and or context. (Or, the presence of paragraphs and paragraphs of LLM output makes the next token predictor dumber? Unsure.)
https://github.com/spott/pi-task-compaction
I’m still playing with it, but it essentially has the model define a region, and mark it with begin_task, and end_task. End_task also requires a fairly hefty summery of the entire region.
It then pulls that entire region and replaces it with the summary. If the model wants it can actually look into the output that was pruned from the conversation, but in practice the summaries have been good enough it hasn’t needed to.
I’ve been using it with 5.6 sol, and even really long sessions have like 6% context used at the end.
I feel like this is kind of subagents, but it doesn't have the clear task -> work -> return result -> die model, rather it's more closely aligned to how I myself multitask, maintaining several long-lived threads of execution, and having to intentionally "forget" about the other ones when it's time to lock in and do real work on the most important one.
this seems to be the fate with all the big grand 'god' projects people are doing with AI. super epic omega knowledge bases, super epic omega personal platforms, etc. I don't think enough people admit that whatever they're working on has evolved into something they don't fully understand.
https://pchalasani.github.io/claude-code-tools/tools/aichat/...
There are other pruning or compaction strategies you can configure too.
You can also use "/btw" to have side tangent conversations with the current context but with the benefit that it will not bloat your current context after the tangent is done.
I wish other harnesses were prioritising this sort of flexibility with context management, but they're likely more incentivised on optimising their cache rates
https://github.com/can1357/oh-my-pi
Ah, something I needed, but did not know I needed. Comes up ALL the time. I want to know something that is well off the task, but it's the best time to ask due to the specifically loaded context.
Sounds cool and it does make sensible decisions optically but I haven’t been able to prove that it is meaningfully better than normal compaction. Building harnesses that do interesting things is a lot easier than building more effective harnesses, I guess.
Having a second model also iterate the resulting messages and remove low-value tool calls could also be interesting. Especially failed calls which add no value.
https://pi.dev/packages/context-fold
That naturally trims a lot of context while removing mistakes from the context to prevent poisoning (ie, every draft but the final negatively contributes in some way — that’s why we edited them).
Because juggler has a nested-thread architecture, I had a great ah-ha! moment when I realised that with sub-threads, compaction becomes almost free: you just take the list of items in the conversation, move them all into a new sub-thread, then allow that sub-thread to summarise itself (this summary is what the parent thread sees). Hey-presto, your parent thread is now compacted, and the sub-thread contains all the old messages, so you can browse them if you need to, or just undo the whole operation if you change you mind. This struck me as super-elegant, and I was chuffed that it works.
However.. What I find myself actually doing now is just using the "new conversation" tool to hand over to a new conversation - I just tell the LLM "create a new conversation, and tell it everything it needs to continue this task". It gets you to the same place, but generally I'm not interested in the history, and having a fresh conversation with a new name usually feels neater.
User:'How is auth implemented?' -> [thinking] [codebase exploration with [thinking] in between, 10 file reads, 3 of which were "wrong"] [thinking] -> agent_response
This little exchange contains a WHAT (how auth actually is implemented) and a HOW (where that info is and how to retrieve it). Maybe this question was part of a larger task. I think that whole exchange could be summarised before it enters context, kind of like what happens with subagents. The main thread would then consist mostly of [summaries]. Eventually the context will fill up anyway and we would summarise those summaries again. Alternatively one could maintain a [master_summary], kind of like an internal state. So new [summaries] get integrated directly and the [master_summary] gets updated.
Here's what I do for each of my sessions:
1. For asides, off-topic work, or repetitive work that has already been done in the session, branch backwards (with /tree) and summarize.
2. If I've exceeded 30% or the 'price-doubling' multi-tier pricing, prune (my custom extension).
3. If I've already pruned and I'm still close to 30%, 'prune all' (more extensive prune).
Definition:
'/prune': Removes ~50% context on a fresh session (not previously pruned)
'/prune-extended': Removes ~80% context on a fresh session Both create a new session and delete the old one after a successful switch.Using these I can keep a session going for weeks (or longer), even with extensive use and almost all the important context is preserved while dumping the less important context. Neither command requires an LLM summarization so they execute quickly.
One trick I find useful is to have one model with two KV caches running and while first cache has produced tokens, second cache immediately summarizes them during input tokens are being generated (tools time), then harness switches to the second KV cache which takes newly produced input tokens while KV in first cache is getting replaced with compacted summary tokens. This is a kind of ping pong, so we trade more space for less time. Still experimenting but it looks it works, and nice bonus it improves GPU utilization. Btw I have my own harness and model serving code, but it can be easily implemented in any other harness and model server.
It's like asking the questions needed to get the answer you need from the LLM, can be either an answer for a plan/todo or a task to code/change something.
> Extensions can intercept and customize both compaction and branch summarization
https://pi.dev/docs/latest/compaction
Just make an extension (or ask Pi to write an extension for itself) that intercepts compaction and leaves only what you want, or rewrites it in any other way. Should be just a few lines.
It sounds crazy, but it's way cheaper and faster for multimodal models to read text on images than the equivalent text represented as tokens. Text tokens take a lot more bytes to represent internally than you would think. Give it images with rasterized text, and there is no OCR-like intermediate step that turns images into text tokens. It just goes straight into the multimodal context where it's all just linea algebra underneath. The cost is more lossiness than if you gave it the same tokens as text. Which is why you don't just render text prompts and files to read as images.
But compaction is inherently lossy, a summary 1-10% of the original length is going to lose things, by definition. So you keep the things you really need to get exactly right in text token summary, but put a lot more of what you would have just discarded into a PNG. Keeping the past context as text in an image is way less lossy than keeping a text summary of that context.
They peeked inside the layers and found the same neurons were usually activating in the middle layers when fed either text tokens or images of that same text. It just takes the model more time and kv cache to do it via text tokens.
Deepseek was the first to really jump on this: https://github.com/deepseek-ai/DeepSeek-OCR/blob/main/DeepSe...
Say, what happens when chain of summaries grows so long, that it still overflows context window. Is summarization runned over the summaries in the context window?
I pretty much just start a new session whenever i fill the context.
For an EPYC with a 5090 (no layers on CPU) vs an M3 max 128GB, qwen 3.6 27B at 128k context / 7k generation:
This is for dense qwen (which I wouldn't run day to day on the mac) - in reality the mac is quite usable with MoEs but you definitely notice a difference.LLMs are perfectly capable of summarising the conversation without a new system prompt.
This way you have a prefix cache hit on the session and don't pay any more than for a regular prompt.
The prompt is assembled as tool schemas, then system instructions, then message history, and caching works on prefixes of that. So the cost of an edit is not its size, it is the size of everything behind it. Edit a tool definition and you have invalidated the system prompt and the whole conversation with it. Edit the tail and you pay for the tail.
I measured this by accident in a multi agent ablation where the only variable was whether the supervisor sent a fixed tool array or a per-task subset to the worker. Runs paying cache creation, out of 120 each: fixed 0 cache creation, per-task subset 58 cache creation. Under a prompt load that was $0.0382 per run against $0.0230, and on a clean context it reversed because there was no prefix worth caching.
Which is the argument for what Pi does here. A pointer is a tail edit. Rewriting is a head edit.
[1] https://ampcode.com/news/handoff
My problem is that when in a loop and it’s calling tools, it won’t check how close it is to the compacting limit until the whole loop returns to you for the next prompt. And so if you have a run that could go for hours, it’s a gamble if you’ll OOM or an interrupting compaction breaks context and stops the loop without continuing.
There’s a few extensions that all try to solve this problem, but I’ve found none actually work :(
Their approach is minimalism. If you need something pi doesn’t have, the recommendation is to ask pi about it and have it build an extension for you, or look for a plugin.
That said, try it. You may find you didn’t need everything in those ecosystems.
Not a shill.. I just really like pi.
https://github.com/can1357/oh-my-pi
That said, from reading the pi subreddit I don't see people referencing these much.
See: https://github.com/earendil-works/pi/blob/main/packages/codi...
I get it into 1M+ routinely on local models with operations between 50k-85k
The pruning does evict cache but not from the start, and only up till the last time you pruned it, as it's just adding messages on top.
but yes, it takes some additional tests and docs to keep it from just becoming hollowed out on tasks; I'd say about 10-20% of the time is just horribly loses what it's doing.