Blog·Learning

Persistent Research Agents Need an Evidence Ledger, Not Just Summaries

By Shardul Mane·2 min read

Persistent Research Agents Need an Evidence Ledger, Not Just Summaries

A research agent needs to preserve what it has read, which claims a source supports, and what remains unresolved. A single evolving summary can hide conflicting evidence and make a later session treat an unverified inference as a finding.

An evidence ledger keeps sources, claims, and conclusions distinguishable. It can be implemented with ordinary application records; the important part is the contract, not a particular storage technology.

Give sources stable identities and versions

Record the source URL or file ID, retrieval time, relevant publication date, revision, and permitted audience. Keep a locator for the passage that supports a claim. “Read the report” is insufficient when the report changes or contains several conflicting statements.

Separate access to the source from possession of a summary. A later user may not be permitted to see the same material, and a stored conclusion can still reveal information from an inaccessible source.

Store claims with their evidentiary status

Use statuses such as proposed, supported, contradicted, and unresolved if they fit the workflow. Preserve the supporting and opposing references rather than overwriting disagreement with a confident synthesis.

For example, one vendor guide may say a feature is available while an older deployment note describes a limitation. The useful record identifies the versions and the question to verify. It does not choose whichever statement was ingested last.

Make handoffs preserve unfinished work

A handoff should state which sources were inspected, which questions remain open, and which actions actually completed. “Draft prepared” must not become “report sent.” Preserve exact figures and source references where paraphrasing would make later verification difficult.

The decision-record guide covers confirmed outcomes. A research ledger also needs room for evidence that has not yet justified a decision.

Recheck sources when the question requires it

“Already read” should reduce duplicate work, not prohibit a fresh fetch. Time-sensitive claims need current verification. Record why a source was revisited and whether the new version changes the conclusion.

Test a conflicting source, a revised document, a permission change, and a new session that must continue an unresolved question. Evaluate whether the agent can explain its uncertainty with evidence.

For the ingestion foundation, see the knowledge-base guide. Try Supermemory with a small fictional research corpus while keeping the ledger and its review states explicit in your application.

  1. An update to supermemoryWe've discontinued the supermemory company brain and Nova. Everyone who was charged has been refunded, our MCP and plugins continue to run, and we're going all in on the memory engine.
  2. Scaling Conversations: How Adapta Grew Usage Without Losing ContextAdapta added Supermemory as a persistent memory layer so every conversation keeps its context — letting the team scale usage without losing the thread.
  3. How Chatarmin Ditched RAG and Went Memory-Only with SupermemoryChatarmin replaced a heavy RAG pipeline with Supermemory's memory layer — cutting average AI response time from 40s to 12s and token usage by 40–50%.
  4. SMFS: making agentic retrieval 55% cheaper AND more accurateWe launched SMFS.ai (Supermemory Filesystem) a few weeks ago, with a simple bet: We can redesign the filesystem specifically for agents, with special files, structures, and commands that it can use for it's tasks. Today, SMFS is used by hundreds of companies to power their agents.
  5. Introducing Dynamic Dreaming: supermemory now connects the dots, for you.Dreaming is magical. TLDR: We're launching Dynamic Dreaming in supermemory today, which automatically works if you're using supermemory in any way - API, OpenClaw, Hermes agent, etc.
  6. Dear reader, we just made supermemory insanely cheap... the Context CloudWhen I first started building supermemory, I had one goal: To build the best memory system for AI. I would talk to customers, and find out that memory was not the only thing they needed - They were all setting up 7-8 different vendors at the same time.
  7. Introducing @supermemory/tools v2.0.0Today we're releasing v2.0.0. This release unifies the API across all agents sdk integrations from AI SDK to Mastra, makes conversation identity a first-class concept, and ships with memory saving on by default.
  8. Solving the Precision-Recall Tradeoff: Search Result AggregationWhen you're building memory for AI, search is your foundational layer. The way search generally works is straightforward: the user defines a query, and then sets a limit (top-K) on how many search results they want returned. Usually, this is set to 10 or 20.
  9. Stateful Coding Agents with Memory: Build Long-Running Agents (2026)We built a plugin for Claude Code and OpenCode that gives your coding agent persistent memory. It remembers your preferences, learns your codebase, and never loses context mid-conversation. The result is an agent you can run for months without starting over.
  10. OpenClaw Memory Problems: Why It Forgets and How to Fix It (2026)TLDR: Today, we are releasing a new version of our openclaw plugin - https://github.com/supermemoryai/openclaw-supermemory. This post is going to be a bit technical, so bear with me (or bookmark for later!) In this post, I will talk about what we do about OpenClaw memory, and how we fix it.
  11. Clawd / Molt bot's memory SUCKS. We gave it supermemory.I'm the founder of supermemory. Clawd/Molt bot is blowing up right now, with many, many use cases. I set it up, too, and have been using it through telegram. TLDR: just go to https://supermemory.ai/docs/integrations/clawdbot to set up supermemory for your clawd bot.
  12. Catch up with our UNFORGETTABLE Launch WeekOver the last year, one belief has guided almost everything we’ve built at Supermemory AI becomes meaningfully useful only when it remembers. Memory shouldn’t be something developers rebuild from scratch. It shouldn’t be fragile, expensive, or trapped inside a single tool.
  13. Empowering the Next Generation of Founders: Supermemory Startup ProgramIf there’s one thing we’ve learned while building Supermemory, it’s that most startups don’t fail because they didn't build features; they fail when infrastructure slows them down, or they built too slow.
  14. Building code-chunk: AST Aware Code ChunkingAt Supermemory, we're building context engineering infrastructure for AI. A huge part of that is dealing with code: ingesting repos, understanding structure, and making it searchable. The problem is that most code chunking solutions are terrible. We built code-chunk to fix this.
  15. Supermemory raises $3 million with the best memory engine for LLMsToday, I am excited to announce our first funding round to accelerate our mission of building an interoperable, scalable and reliable memory for LLMs and agents. Memory is one of the hardest challenges in AI right now.
  16. Mem0 vs Supermemory: Why Scira SwitchedScira AI moved its production memory layer from Mem0 to Supermemory. This is what failed, what improved, and how the team evaluated the two systems.
  17. Never Record Again: How Montra Uses Supermemory to Rethink Video CreationCampbell Baron, the founder of Montra, has been making videos since he was twelve. By thirteen, he was already doing brand work. Today, he’s betting on a very different future for creators: a world where recording is the exception, and most videos are generated from scratch.
  18. Unified Memory That Works Where You Work: Your Second Brain With SupermemoryHi everyone, I’m Dhravya, the founder of Supermemory. I want to start with a little story behind why this product means so much to me. You can also skip straight to what it is and how it works below.
  19. Supermemory just got faster on PlanetScaleWhat is Supermemory? Supermemory completes the missing part of the LLM puzzle: memory. Just as memory is crucial for human intelligence, it's essential for truly intelligent AI systems.
  20. Faster, smarter, reliable infinite chat: Supermemory IS context engineering.People are obsessed with prompts and prompt engineering. Sure, what you say is important, but what the model knows when you say it is the difference between a stateless text generator and an intelligent AI system. In short, context is the most crucial component.
  21. We solved AI API interoperabilityOne API to rule them all, One spec to find them, One library to bring them all and in the TypeScript, bind them. When we were building the the Infinite Chat API, initially, we only supported the OpenAI format. This was fine, until a lot of our customers started asking, asking for more.
  22. The Wow Factor of Memory - How Flow Used Supermemory To Build Smarter, Stickier ProductsOverview: Flow is a note-taking app built around a bold vision: to create a more personal, context-aware writing experience powered by AI. At the heart of this mission is memory.
  23. The UX and technicalities of awesome MCPsLast month, we launched the Supermemory MCP, mostly to test our own infrastructure and get some initial traction. It blew up. To my absolute surprise, the initial launch itself got half a million impressions (!!!). Then, we launched and got #2 on ProductHunt too.
  24. Architecting a memory engine inspired by the human brainLanguage is at the heart of intelligence, but what truly powers meaningful interaction is memory — the ability to accumulate, recall, and contextualize information over time. Large Language Models (LLMs) have mastered language, but memory remains their Achilles’ heel.