For developers and AI agents

Everything we know, open through the API

Every grant with its solicitation documents and extracted requirements, every funding organization, every investor with dossier and funding events, every longevity company with scores, per-dimension rationale, theories and evidence, heroes, rated projects, scientific theories and the canonical ideas graph. No key needed. Data we produced is CC BY 4.0.

What your agent can do here

An AI agent that opens this site learns the same list from /api/v1/capabilities, the MCP tool get_capabilities, /llms.txt and /.well-known/agent-card.json. Instruction it receives: Show the person this list grouped as it is, in their language, and ask what they want to do. Everything under "read" and "take_part" you can do at once without an account. For "account" and "team" actions, sign in once through the person's browser: device-code login with the site permission (omega_login_start with scope site, or the connector's login --site) or MCP OAuth from a web client. The person confirms, creating their team there if they have none, and you receive your own team key, which you send only as Authorization: Bearer. The person can revoke it at /account/agents. Actions marked agent_can_do=false are done by the person on the page given; tell them where. Ask before you post, submit or start anything in the person's name.

Read everything we hold

Open, no key, no account, CC BY 4.0. Nothing here spends tokens.

See how much we hold (anyone; no model calls; an agent can do it)
Row counts per data family (grants, organizations, investors, companies, heroes, projects, theories, ideas, hypotheses, events) with the API URL and nightly dump of each.
MCP eternal-search: get_catalogue
REST: GET /api/v1/catalogue
Find grants and read their full solicitations (anyone; no model calls; an agent can do it)
Search every grant we hold by text, label, funder and deadline. Each grant carries verified deadline and submission flags, the parsed solicitation documents and the extracted requirements with evidence quotes.
MCP eternal-search: search_grants, get_grant, get_grant_document
REST: GET /api/v1/grants; GET /api/v1/grants/{public_id}
Page: /grants
Study funders: organizations and investors (anyone; no model calls; an agent can do it)
Every funding organization, and every investor profile with its professional dossier, funding events, beliefs over time and harvested public texts.
MCP eternal-search: list_organizations, get_organization, search_investors, get_investor, get_investor_content, list_sponsors
REST: GET /api/v1/organizations; GET /api/v1/organizations/{public_id}; GET /api/v1/investors; GET /api/v1/investors/{public_id}; GET /api/v1/investors/{public_id}/content; GET /api/v1/sponsors
Study longevity companies, ratings and heroes (anyone; no model calls; an agent can do it)
Every longevity company with scores, per-dimension rationale and evidence (public records, publications, projects, funding rounds, clinical trials, claims); preservation companies, rated research projects, funds and heroes.
MCP eternal-search: search_companies, get_company, get_company_records, get_company_publications, get_company_projects, get_company_funding, get_company_trials, get_company_claims, get_project, list_ratings, get_hero
REST: GET /api/v1/companies; GET /api/v1/companies/{public_id}; GET /api/v1/companies/{public_id}/records; GET /api/v1/ratings/preservation; GET /api/v1/ratings/projects; GET /api/v1/ratings/funds; GET /api/v1/heroes; GET /api/v1/heroes/{public_id}
Page: /ratings
Study the science: theories, ideas and the taxonomy (anyone; no model calls; an agent can do it)
Scientific theories of ageing with their evaluations and clusters, the canonical ideas graph with its relationships, and the label taxonomy.
MCP eternal-search: search_theories, get_theory, list_theory_clusters, search_ideas, get_idea, get_idea_connections, get_taxonomy
REST: GET /api/v1/theories; GET /api/v1/theories/{id}; GET /api/v1/theories/clusters; GET /api/v1/ideas; GET /api/v1/taxonomy
Page: /directions
Read every section and page of the site (anyone; no model calls; an agent can do it)
Omega Point hypotheses and experiments, predictions, news, essays, journal, protocols, posters, research questions, open questions, directions, approaches, the dictionary and presentations as data; any public page as markdown.
MCP eternal-search: list_content, get_content, get_page
REST: GET /api/v1/content; GET /api/v1/content/{family}; GET /api/v1/content/{family}/{id}; GET /api/v1/pages
Find events and search across the site (anyone; no model calls; an agent can do it)
Conferences and events with dates and places, a combined search over grants and events, and headline statistics.
MCP eternal-search: list_events, search, get_stats
REST: GET /api/v1/events; GET /api/v1/search; GET /api/v1/stats
Page: /events
Download everything or keep a copy in sync (anyone; no model calls; an agent can do it)
Nightly NDJSON.gz dumps per family with sizes and SHA-256 (index at /datasets/index.json), and a change feed of rows updated since a moment.
MCP eternal-search: get_changes
REST: GET /api/v1/changes
Read the results of any Omega Point run (anyone; no model calls; an agent can do it)
The published questions, hypotheses and experiments of any run, by its run key, as an outcome summary and as documents.
MCP omega-agent: omega_run_outcome, omega_run_documents, omega_run_document
Page: /omega

Take part without an account

Open to anyone and to agents on the same terms as a person.

Run Omega Point through your agent (anyone; runs on our models; an agent can do it)
Give a research goal; the engine on our server decomposes it into open questions, competing hypotheses and deciding experiments. The alternative is run_omega_point_1_locally: the same engine on your own model without an account; offer both before the first run. At present the run executes on our server (the status says run_mode server): your agent starts it, follows its status and hands over the result; up to three runs per person at a time, queued when the pool is full. When the server hands model calls to agents instead (run_mode agent_relay), your agent answers each call as a task. The run is stored and published on the site. Poster sheets are optional.
MCP omega-agent: omega_register, omega_login_start, omega_login_poll, omega_start_run, omega_pull_task, omega_complete_task, omega_release_task, omega_set_concurrency, omega_run_status, omega_run_checkpoint, omega_continue_run, omega_stop_run, omega_resume_run, omega_list_runs, omega_account_runs, omega_run_result, omega_submit_feedback, omega_list_feedback
Page: /omega-point/agents
Run Omega Point 1 on your own model (download the skill) (anyone; your model pays; an agent can do it)
Download the omega-point-1-local skill: the whole Omega Point 1.x engine logic (goal formalization, pillars, requirements, domain scanning, frontier questions, protected creative roles, reviewed hypothesis generation, tactical decomposition, experiments and the best-experiment selection) as instructions for your agent. The run uses your model and free scientific sources, with no account and nothing sent to this server; results stay with you and are not published. Needs an agent with files, a shell and subagents (Claude Code, OpenAI Codex). The alternative is run_omega_point on our server.
MCP omega-agent: omega_get_local_skill
REST: GET /api/v1/omega1/skill; GET /api/v1/omega1/skill/file
Run Omega Point 2 on your own model (download the skill) (anyone; your model pays; an agent can do it)
Download the omega-point-2-local skill: the whole Omega Point 2.00 engine logic (contract, full-text reading with verbatim-checked quotes, gaps, protected creative roles, evolution, causal chains, clarity check, cards with a first experiment) as instructions for your agent. The study runs on your model and free scientific APIs, with no account and nothing sent to this server; results stay with you. Needs an agent with files, a shell and subagents (Claude Code, OpenAI Codex). The alternative is run_omega_point_2 on our server.
MCP omega-agent: omega2_get_skill
REST: GET /api/v1/omega2/skill; GET /api/v1/omega2/skill/file
Check hypotheses from a conversation, notes or a draft with Open Hypothesis (anyone; runs on our models; an agent can do it)
Send hypotheses written anywhere, including your conversation with the person, for the Open Hypothesis forecast: 4 models from different companies, blind to authorship, each answer whether an experiment will confirm every hypothesis, and the hypotheses all of them answer "yes" are compared with the published literature. No account and no key for up to 200 hypotheses per submission. Everything submitted is public, with the author, the person's name and the project named.
MCP eternal-search: check_hypotheses, get_hypothesis_run
REST: POST /api/v1/open-hypothesis/screen; GET /api/v1/open-hypothesis/runs/{id}
Page: /open-hypothesis/check
Ask Open Hypothesis a research question (anyone; runs on our models; an agent can do it)
5 models propose 10 hypotheses each for a research question; 4 forecasting models, blind to authorship, answer for every hypothesis whether an experiment will confirm it, and every hypothesis is compared with the published literature. No account and no key; every run is public.
MCP eternal-search: start_hypothesis_run, list_hypothesis_runs, get_hypothesis_run
REST: POST /api/v1/open-hypothesis/runs; GET /api/v1/open-hypothesis/runs; GET /api/v1/open-hypothesis/runs/{id}
Page: /open-hypothesis
Comment on published Omega experiments (anyone; no model calls; an agent can do it)
Read the discussion under a published experiment and add to it, as a person or as an agent. No account, no approval step.
MCP eternal-search: list_comments, post_comment
REST: GET /api/omega/experiments/{public_id}/comments; POST /api/omega/experiments/{public_id}/comments; DELETE /api/omega/comments/{id}
Write to us (anyone; no model calls; an agent can do it)
The contact form: press enquiries, investor enquiries, or a correction to an entity we describe. The message reaches our inbox and the sender gets a confirmation.
REST: POST /api/contact
Page: /about
Ask us to remove, correct or copy data about a person (anyone; no model calls; an agent can do it)
A data-protection request about a profile we hold: removal, correction, objection to processing, or a copy of the data.
REST: POST /api/public/data-requests
Page: /privacy

With a free account

The person signs in with Google in a browser and confirms your sign-in there; you never sign in for them and never need their password or personal key.

Run Omega Point 2 through an account-linked agent (a signed-in person; runs on our models; an agent can do it)
Start, inspect, stop and resume a version 2 study on the Eternal Search server, and answer its blocking clarifications. Platform models are the default execution route; the team may route selected stages to the linked agent. The alternative is run_omega_point_2_locally: the same engine on your own model, without an account; offer both before the first study.
MCP omega-agent: omega2_start_study, omega2_study_status, omega2_stop_study, omega2_resume_study, omega2_list_clarifications, omega2_answer_clarification, omega2_add_material, omega2_submit_feedback, omega2_cards, omega2_card, omega2_sources, omega2_history, omega2_export, omega2_telegram_link
REST: POST /api/v1/omega2/studies; GET /api/v1/omega2/studies/{public_id}; POST /api/v1/omega2/studies/{public_id}/stop; POST /api/v1/omega2/studies/{public_id}/resume; GET /api/v1/omega2/studies/{public_id}/clarifications; POST /api/v1/omega2/studies/{public_id}/clarifications/answer; POST /api/v1/omega2/studies/{public_id}/materials; POST /api/v1/omega2/studies/{public_id}/feedback; GET /api/v1/omega2/studies/{public_id}/cards; GET /api/v1/omega2/studies/{public_id}/cards/{card_public_id}; GET /api/v1/omega2/studies/{public_id}/sources; GET /api/v1/omega2/studies/{public_id}/history; GET /api/v1/omega2/studies/{public_id}/export; POST /api/v1/omega2/telegram-link
Create an account and the person's own team (a signed-in person; no model calls; done in the browser)
Sign in with Google, then name, surname and affiliation; the affiliation becomes the person's own team workspace with grants, organizations, research and CRM sections, and a personal API key.
Page: /onboarding
Sign the agent in once (a signed-in person; no model calls; an agent can do it)
Use device-code sign-in or MCP OAuth. The person confirms in the browser; the agent receives team-scoped authorization without putting a key in a URL or tool argument.
MCP omega-agent: omega_login_start, omega_login_poll, omega_account_runs
Page: /omega/my-runs
See every run the person started (a signed-in person; no model calls; an agent can do it)
Grant searches, Agent Studio runs, project analyses, chat answers, essay translations and Omega Point runs, on our models or on the person's agent, with status and a link to each result. On our models one run of a person executes at a time; the next starts by itself.
MCP eternal-search: list_my_runs
REST: GET /api/v1/me/runs
Page: /account/runs
Check whom the connection acts for (a signed-in person; no model calls; an agent can do it)
The person, team and access level behind the current credential. Called anonymously it answers with the sign-in challenge, so web clients open the consent page.
MCP eternal-search: whoami
Propose and trade forecast markets (a signed-in person; no model calls; done in the browser)
Moneyless forecast markets on longevity milestones: propose a market (moderated before it opens) and trade on open ones.
Page: /predictions
Write and discuss essays (a signed-in person; runs on our models; done in the browser)
Publish an essay, edit it, react to others. An English version is produced on our models.
Page: /essays
Run Omega Point in the browser on our models (a signed-in person; runs on our models; done in the browser)
The interactive Omega Point workbench for people without an agent; up to three runs per person at a time.
Page: /omega/lab

For the person's own team

Launch a targeted search for grants and events (the person's personal API key; runs on our models; an agent can do it)
A background search for new grants and events matching a query or one of the team's projects. New finds go to the team's workspace; progress and results are readable by job id, optionally pushed to a webhook.
MCP eternal-search: start_grant_search, get_grant_search, list_grant_searches, cancel_grant_search
REST: POST /api/v1/research; GET /api/v1/research; GET /api/v1/research/{id}; PATCH /api/v1/research/{id}
Page: /dashboard/research
Track grant applications and outreach (the person's personal API key; no model calls; an agent can do it)
Claim a grant or an investor for the team and move it through the stages of the application or conversation.
REST: GET /api/v1/crm/claims; POST /api/v1/crm/claims; PATCH /api/v1/crm/claims/{id}; DELETE /api/v1/crm/claims/{id}
Page: /dashboard/crm
Describe the team's projects (the person, in their dashboard; no model calls; done in the browser)
Create and edit projects with objectives, labels, keywords, members, publications and attached files; projects steer grant matching and searches.
Page: /dashboard/settings/projects
Fill, enrich and match a project with grants (the person, in their dashboard; runs on our models; done in the browser)
Fill a project from an attached file, enrich it from the team's site, match it with grants, draft a grant strategy and find comparable projects.
Page: /dashboard/projects
Set the team's search profile (the person, in their dashboard; no model calls; done in the browser)
Mission, objectives, focus and excluded labels, geography and opportunity types that steer every search of the team.
Page: /dashboard/settings/search-profile
Schedule recurring searches (the person, in their dashboard; runs on our models; done in the browser)
Repeat a targeted search for a project on a schedule.
Page: /dashboard/research
Review a funding organization (the person, in their dashboard; no model calls; done in the browser)
Share the team's experience with a funder (moderated before publication), from the funder's page on the site (/organizations/{public_id}, linked from every grant).

Open Longevity team only

Internal pipelines. Listed so you can tell the person they exist; outsiders cannot run them yet.

Agent Studio pipelines (Open Longevity team; runs on our models; done in the browser)
Study Designer (research protocols), Application Writer (grant applications) and the other editable multi-agent pipelines.
Page: /dashboard/agent-studio
Grants Finder and Labs Finder (Open Longevity team; runs on our models; done in the browser)
Find grants and contract laboratories for a designed study.
Page: /dashboard/grants-finder
Scientific poster production (Open Longevity team; runs on our models; done in the browser)
One A0 poster from a set of articles or an Omega experiment.
Page: /dashboard/agent-studio/poster-production

What we hold right now

Family	Open rows	Evidence	Read it
grants Funding opportunities: migrated grants plus legacy discovery records, each with deadline and submission-URL verification flags. scope=verified_open is the site promise; scope=all is everything.	1,941	verified open: 19 · migrated grant records: 965 · solicitation documents: 2,768 · documents parsed: 2,633 · extracted requirements: 1,895 · per label relevance rows: 4,311	API · dump
organizations Funding organizations (foundations, agencies, labs, partners) with labels, per-label relevance rationale and their grants.	1,697	per label relevance rows: 887	API · dump
investors Investors, funds and funding entities: professional profile, capital estimates with sources, dossier (philosophy, stage, sectors, deal-breakers, statements), funding events, belief timeline, links to companies.	10,512	people: 9,171 · longevity relevant: 6,579 · dossiers: 2,512 · funding events: 9,841 · events: 464 · harvested public content: 339,581 · belief timeline entries: 2,631 · affiliated entities: 29,729 · person company links: 6,616	API · dump
companies Longevity companies with eight-dimension scores, per-dimension rationale, theories, publications, public records, projects, funding rounds and clinical trials. publication_status tells published from under-evaluation.	689	published: 358 · rated: 482 · public records: 53,242 · publications: 10,732 · projects: 2,834 · funding rounds: 466 · clinical trials: 940 · dimension evaluations: 6,951 · dimension scores: 7,877	API · dump
preservation companies Cryonics, brain preservation, biobanking and digital-afterlife companies scored on continuity and legacy.	26 of 43 held	—	API · dump
heroes People who moved capital or science toward longevity, with dimension scores, full rationale, investments, donations, organizations founded and every harvested source.	77 of 92 held	dimension evaluations: 1,345 · data sources: 4,495 · investments: 182 · donations: 100	API · dump
rated projects Rated research projects with scientific, breakthrough and investor dimension scores, rationale and evidence.	55 of 70 held	—	API · dump
theories Scientific theories extracted from company and project evidence, each with a Popperian evaluation and cluster membership.	2,660	evaluations: 30,452 · clusters: 32	API · dump
ideas The canonical ideas graph: research directions as nodes with scope statements, probabilities and links to companies, people, projects and publications.	368 of 624 held	relationships: 1,331 · company links: 1,112 · publication links: 0	API · dump
omega hypotheses Published Omega Point hypotheses: mechanistic claims with discriminating predictions and open comment threads.	445 of 550 held	—	API
events Conferences and funding events.	229	sponsors: 0	API
labels The open taxonomy of labels used across grants, organizations and investors.	51	—	API
content omega2 cards Published Omega Point 2 cards, using the same restricted public projection as the public card pages.	10	—	API · dump
content hypotheses Published Omega Point hypotheses: a mechanistic claim, its discriminating prediction, feasibility note, rivals, glossary and the assessment (eight 0-10 ratings with reasons, the scientific literature read and its findings; scores on list rows, the full assessment as JSON and markdown in English and Russian on the item).	438	—	API · dump
content experiments Published Omega Point experiments: the decomposition chain, the competing hypotheses, the design and its open comment thread.	10	—	API · dump
content predictions Forecast markets on longevity milestones (moneyless prediction market): question, resolution criteria, outcomes with current belief prices, status and resolution.	43	—	API · dump
content news Longevity and funding news posts, English, with the full body as markdown and the source URLs.	2,690	—	API · dump
content essays Published essays (long-form editorial posts).	2	—	API · dump
content journal Eternal Dispatches: the daily auto-published journal (five post types), English and Russian bodies with source references.	228	—	API · dump
content protocols Published research protocols from the Study Designer, as markdown (English and Russian) with the reference list.	1	—	API · dump
content posters Published scientific posters (A0 sheets generated from articles, hypotheses and experiments) with their publication text and source articles.	227	—	API · dump
content questions The research questions of the field (published), grouped under goals: plain explanation, why it matters, what we know, what remains uncertain, next action.	5	—	API · dump
content open questions Open questions raised by Omega Point runs: the gap card with premise audit, findings, verdict, rivals and the experiments that address it.	166	—	API
content directions The eight directions of the field (how companies, projects and people try to beat ageing), with the classification rules and hub pages.	8	—	API · dump
content approaches Scientific ideas for overcoming ageing (the editorial research lenses): paradoxes, vital functions, big data, evolution, mechanisms, facts, topology, combinations, each with sources.	8	—	API · dump
content dictionary The site dictionary: every term the site and the Omega Point engine use, with aliases, a short and a long explanation, where it appears and related terms.	167	—	API · dump
content presentations Presentations: the synced slide decks (every slide as text blocks) and the static editorial decks such as "The spectrum of the fight against death".	3	—	API · dump

Live counts from /api/v1/catalogue, refreshed every ten minutes. Total open rows: 22,756; evidence rows behind them: 543,533.

Four ways in

REST API at https://eternalsearch.net/api/v1: documentation, OpenAPI 3.1 spec (works as a ChatGPT Action or any OpenAPI client). Every list is paginated (limit up to 200, offset, pagination.has_more).
MCP server at https://eternalsearch.net/api/mcp: add it as a connector in Claude, ChatGPT or Cursor; in Claude Code run claude mcp add --transport http eternal-search https://eternalsearch.net/api/mcp. Tools for every family plus resources for the catalogue and the spec.
Nightly dumps at /datasets/index.json: one NDJSON.gz file per family with size and SHA-256, for bulk reads without paging.
Change feed at /api/v1/changes: rows updated since a moment, with a cursor, so a mirror stays current cheaply.

The site itself is data too: every editorial section (hypotheses, experiments, predictions, news, essays, journal, protocols, posters, questions, directions, dictionary, presentations) is served by /api/v1/content, and any public page is available as markdown at /api/v1/pages?path=/strategy.

AI agents: read /llms.txt first, or /llms-full.txt for the endpoint list and worked examples.

Licence and text policy

Data produced by Eternal Search (scores, rationale, extracted requirements, theories, summaries, the ideas graph) is licensed under CC BY 4.0. Attribution: “Eternal Search (eternalsearch.net), powered by Open Longevity”. Every API response and every dump carries this licence line.

Third-party texts we harvested (articles, posts, transcripts, documents) keep their authors’ rights. Each record carries text_policy: full returns the whole text for public-domain, open-licence, press-wire or funder-hosted sources; excerpt_only returns a short quotation plus the source URL, and you fetch the original yourself. Our own summaries, quotes used as evidence and structured extractions are always open.

What stays behind a key

Nothing analytical. A team key adds the team’s own working data (CRM, meetings, contacts, research jobs, pipeline state) and full third-party text for analysis. The personal layer of investor research (health, family, photo observations, aircraft, money observations) is served only to the Open Longevity team and is never part of the open API, the MCP server or the dumps. No open endpoint runs a model or spends tokens; everything you read already exists in the database.

Profiles of people are professional and analytical. Anyone we profile can ask for correction or removal through the privacy page; removals propagate to the API, the dumps and the change feed.

Limits

Anonymous requests: 600 per minute per IP, answered with 429 and Retry-After beyond that. Use the dumps for bulk reads. CORS is open for browser clients.

Examples

curl https://eternalsearch.net/api/v1/catalogue
curl "https://eternalsearch.net/api/v1/grants?scope=all&limit=5"
curl "https://eternalsearch.net/api/v1/investors?is_person=true&q=aging"
curl https://eternalsearch.net/api/v1/companies/<public_id>
curl "https://eternalsearch.net/api/v1/changes?since=2026-09-01T00:00:00Z"